Generative AI has rapidly progressed from chatbots and assistants to agents across diverse applications and domains. Generative AI agents demonstrate sophisticated operation through autonomy, tool use and decision making with minimal human input. Despite these performance gains, agents are still impacted by the foundational limitations of Generative AI models. Among these, hallucinations are a major limitation that affects agent operation in real-world settings, leading to risk and loss. Several recent work aim to address hallucinations through methods such as retrieval-augmented generation and reflection prompting, however, these only provide partial improvements. An effective yet underexplored approach is in the observability data generated by an agent in its deployed and operational settings. Drawing on agent observability data, this paper proposes a dual memory knowledge graph approach that integrates Semantic and Observability Memory to address hallucinations in Generative AI agents. Semantic Memory provides organized domain knowledge for precise factual grounding. Observability Memory transforms logs, traces, and execution results into agent validated planning histories. Hallucinations are then addressed by grounded planning in verified past interactions with known, reliable outcomes. This approach is evaluated in a two-stage experimental setup aligned with its dual memory design. Observability memory is evaluated on the HotpotQA dataset to assess its impact on reasoning grounding, using metrics that capture both factual accuracy and reasoning hallucinations. The SM3-Text-to-Query benchmark and Synthea-based medical QA datasets are used to assess factual grounding of the semantic memory. Results from both experiments demonstrate reductions in hallucinations, with semantic memory for contextual grounding reducing factual hallucinations, and observability memory for reasoning grounding reducing faithfulness hallucinations.

With Cyber-Physical Systems (CPS) ranging from smart grids and self-driving cars to healthcare networks becoming the cornerstones of infrastructure in today’s world, detecting anomalies in real-time is crucial to ensure safety, guarantee reliability, and preserve performance. Traditional supervised learning methods face challenges in real-world CPS due to the limited availability of labeled data. Even when labels exist, they often become outdated because of sensor drift, system reconfigurations, or new types of attacks. This survey presents a comprehensive review of recent self-supervised learning (SSL) methods for anomaly detection in CPS, categorized into contrastive, reconstructive, predictive, and joint embedding-based paradigms. We further summarize hybrid SSL models that integrate SSL with generative modeling, domain adaptation, federated learning, and meta-learning to enhance adaptability and deployment feasibility. Drawing from reported results on real-world CPS datasets, we benchmark 29 methods using standard metrics such as F1-score and AUROC, while noting the limitations of commonly used CPS datasets. To the best of our knowledge, this is the first survey that (1) systematically reviews both traditional and hybrid SSL approaches tailored to CPS, (2) identifies Joint Embedding Predictive Architectures (JEPA) as an emerging fourth category alongside contrastive, reconstructive, and predictive methods—an inclusion often overlooked in traditional SSL categorizations, and (3) benchmarks recent SSL methods across a wide range of CPS datasets. These contributions offer a unified foundation for researchers and practitioners aiming to develop robust, adaptive, and label-efficient anomaly detection systems in evolving CPS environments.

Solar photovoltaics are on track to becoming the largest renewable energy source by 2029. This means a rapid increase in the number of solar energy generation installations from residential roof-top systems to utility-scale power plants. The current industrial approaches towards predictive maintenance will be insufficient to manage and maintain the increasing numbers of such installations at peak performance. In this paper, we propose hyper-predictive maintenance as a novel approach based on Generative Artificial Intelligence (AI) agents for highly autonomous management of solar energy infrastructure. The proposed Agentic AI framework deploys multiple agents for baseline generation from solar installations, predictive model development, degradation estimation, degradation evaluation and predictive maintenance that combines baseline performance with contextual information to predict faults and potential causes. This framework is empirically evaluated in the real-world solar energy systems of a multi-campus tertiary education institution. The results of these experiments confirm the robust and accelerated performance of Generative AI agents for the hyper predictive maintenance of large-scale solar energy installations.

Ground segmentation from LiDAR point cloud data plays a critical role in both civil engineering and autonomous vehicle systems. However, real-world LiDAR data often suffers from geometric distortions, occlusions, and dense clutter, which limit the reliability and accuracy of ground segmentation. To overcome these challenges, we introduce KPU-Net. This deep neural network architecture employs: 1) T-Net module, which handles geometric distortions by aligning point clouds into a canonical pose, 2) KPConv-augmented U-Net encoder-decoder, which handles occlusions, clutter, and irregular terrain by capturing fine-grained, hierarchical features through learned kernel point convolutions over local neighborhoods. In addition to above, KPU-Net offers following advantages: 3) KPU-Net features high speed processing (approximately 231K points per second making it well-suited for scalable deployment in mapping and perception systems), while 4) preserving the original point cloud density (i.e no loss in point cloud data, avoiding sparsification that can compromise precision in various applications). The framework was trained and tested on a benchmark dataset and diverse point cloud data collected by the Timmons group, covering urban, vegetation, and complex terrain environments. The presented KPU-Net was evaluated against five widely used LIDAR data segmentation methods: Random Forest, PointNet, GndNet, RandLA-Net, and KPConv. The proposed KPU-Net demonstrated better performance on mean Intersection over Union (mIoU, up to 33%), mean accuracy (mAcc, up to 25%), and overall accuracy (OA, up to 19%), over the five compared state-of-the-art methods.

The Internet of Vehicles (IoV) has transformed intelligent transportation systems through vehicle-to-everything (V2X) communication, improving road safety and traffic efficiency. However, the dynamic nature of vehicular networks, with high mobility and shared wireless resources, makes them vulnerable to attacks like Denial of Service (DoS). Anomaly detection (AD) has proven effective in detecting such threats. Yet, V2X communication occurs in diverse environments with varying network coverage and vehicle speeds, leading to domain shifts and variations in feature distributions that can hinder the generalization performance of traditional anomaly detection models. To address these challenges, this paper presents V2XFormer, an unsupervised anomaly detection system based on transformer neural networks, designed to identify anomalies in V2X communication. Additionally, we introduce TV2XFormer, which integrates transfer learning to enhance adaptability across diverse network conditions and environmental variations in V2X communication. We assess the performance of the proposed approaches using the VDoS-LRS V2X dataset, employing precision, recall, and$\mathbf{F 1}$score metrics. A comparison is made with five state-of-the-art unsupervised AD algorithms. Experimental results demonstrate that both V2XFormer and TV2XFormer outperform the competing algorithms, achieving the highest$\mathbf{F 1}$scores (1.0). Furthermore, TV2XFormer exhibits notable robustness and generalizability to dynamic vehicular environments.

Machine learning and deep neural networks (DNNs) have been proposed as a tool to identify anomalies in computer network communications. However, due the obfuscatednature of off-the-shelf machine learning models, their output often does not provide enough information to isolate the source of the anomaly to take corrective measures. In this article, we introduce the network transformer (NeT), a DNN model for anomaly detection that incorporates the graph structure of the communication network in order to improve interpretability. The presented approach has the following advantages: first, enhanced interpretability by incorporating the graph structure of computer networks; second, provides a hierarchical set of features that enables analysis at different levels of granularity; second, self-supervised training that does not require labeled data. The NeT model was evaluated on a set of anomalous scenarios executed in a real industrial control system. The presented approach successfully identified the anomalies, the devices affected, and the specific connections causing the anomalies, providing a data-driven hierarchical approach to analyze the behavior of a cyber network.

As Deep Learning (DL) algorithms become more widely adopted in healthcare applications, there is a greater emphasis on understanding and addressing the potential privacy risks associated with these models. The purpose of this study is to investigate the privacy vulnerabilities of the Convolutional Neural Network (CNN) classifiers for Electroencephalogram (EEG) data in the Brain-Computer Interfaces (BCIs). Specifically, it focuses on the Membership Inference Attack (MIA), which seeks to determine if data from an individual were used in model training. The novelty of this work lies in its empirical analysis of MIA, by addressing two key challenges that are less common in other domains: 1) heterogeneous datasets and 2) spatio-temporal design choices. Motivated by these challenges, we investigate the susceptibility to MIA based on: 1) the specifics of the training data set (number of participants, demographics), and 2) specifics of the CNN (such as architecture, regularization). Our experiments revealed that an adversary with limited knowledge of the model and its training process can compromise the privacy of training participants, noting that the same attack is not effective against deep learning models trained on image and tabular datasets. Some of our findings are: 1) training on diverse participant datasets improves the privacy of most participants but increases risks of memorization and vulnerabilities for underrepresented groups; 2) regularization is less effective in defending against the MIA on EEG data CNN classifiers when compared to other types of input data; 3) the depth and width of the model architecture have no impact on the effectiveness of membership attack. We hope that the insights presented will help future researchers develop more privacy-aware deep learning-based BCI systems.

The advancements in Cyber-Physical Systems (CPSs) have also increased their vulnerability to various cyber-attacks. Therefore, it is crucial to develop strong cybersecurity mechanisms, shielding these critical systems from potential cyber intrusions. Among many AI technologies, Generative AI (GenAI) has gained significant attention in the last couple of years. This is due to its distinctive capability to autonomously generate original and diverse content across different domains, offering potential for novel advancements in several applications. Given the massive success of GenAI, it is essential to explore its role in ensuring the cybersecurity of CPSs. Therefore, in this paper, we present: 1) the evolution and current state of GenAI, 2) benefits of GenAI on the cybersecurity of CPS, 3) threats of GenAI on the cybersecurity of CPS, 4) defense strategies against threats and 5) future research opportunities. We hope this systematic survey will help the community prioritize research efforts to address pressing issues in cybersecurity of CPSs.

Large Language Models (LLMs) are leading the Generative Artificial Intelligence transformation in natural language understanding. Beyond language understanding, LLMs have demonstrated capabilities in reasoning tasks, including commonsense, logical, and mathematical reasoning. However, their proficiency in causal understanding has been limited due to the complex nature of causal reasoning. Several recent studies have discussed the role of external causal models for improved causal understanding. Building on the success of Retrieval-Augmented Generation (RAG) for factual reasoning in LLMs, this paper introduces a novel approach that utilizes Causal Graphs as external sources for establishing causal relationships between complex vectors. This method is empirically evaluated using two benchmark datasets across the metrics of Context Relevance, Answer Relevance, and Grounding, in its ability to retrieve relevant context with causal alignment. The retrieval effectiveness is further compared with traditional RAG methods that are based on semantic proximity.

With the many advancements in automobile technology, there has been a sharp increase in the number of sensors and systems present in vehicles. This has enabled a rapid increase in the features and capabilities available in modern automobiles but has also vastly increased their vulnerability surface. Attackers are now able to remotely attack and control some facets of modern automobiles, which creates dangerous situations for drivers and passengers. In order to address this, this paper proposes a self-attention bottleneck network utilizing an encoder-decoder architecture. This is used as an anomaly detection system (ADS) that can detect anomalous behavior present within CAN bus communications. We evaluated this approach using a publicly available CAN bus car hacking dataset and show that our architecture is able to achieve an accuracy of over 99% for detecting anomalies present in CAN bus data.

Electric Vehicle Charging Stations (EVCS) are rapidly being built all around the world to support the growing number of Electric vehicles (EVs) on the road. EVCSs hook into critical infrastructure and offer a vital service, so it is very important they remain secure and available, yet they remain vulnerable to a number of attacks. Thus, it is becoming more and more important to examine the security of EVCSs. In this paper, we will explore the current cyber-physical threat landscape faced by EVCSs. First, we examine the protocols used for communication during the EV charging process and their strengths and weaknesses. Next, we discuss the overall cyber-physical security threats of the entire system, as well as Artificial Intelligence-based (AI) solutions to combat these threats. Finally, we present the future research directions.

With the advancements of modern vehicle infrastructures, vehicles are increasingly relying on the signals received from a vast number of sensors and electronic components. Wireless technologies enable communication between vehicles and infrastructure, but it also increase the vulnerability surface. Malicious actors can remotely disrupt the vehicle's normal behavior, causing vehicle damage or worse, putting human lives in danger. To address these challenges, this paper proposes a transformer neural network-based intrusion detection system (CAN-Former IDS) that predicts anomalous behavior within the CAN protocol communication. Previous work typically addresses the prediction over the sequence of the CAN IDs. In this paper, we will simultaneously analyze both the sequence of IDs and the message payload values. The advantages of our approach are: 1) fully self-supervised training, which does not require labeled data, 2) self learning interactions between input tokens without relying on hand-crafted features. The transformer neural network is trained to predict the next communication sequence and anomalous communication is identified by comparing the real sequence to the predicted expected sequence. We evaluated our approach using a publicly available data set known as survival analysis data set, containing CAN communication from three different cars.

Modern vehicles have multiple electronic control units (ECUs) that are connected as part of a complex cyber-physical system (CPS). The controller area network (CAN) is a well-known communication protocol that connects these ECUs because of its reliability and efficiency. However, adversaries can easily inject abnormal messages into the CAN bus remotely to affect vehicle driving safety. Existing anomaly detection methods only focus on specific vehicle models and have a limited range of applications across different vehicles. To address this challenge, this paper proposes a Domain Adversarial training-based AutoEncoder (DAdAE) for unsupervised CAN anomaly detection. The advantages of our approach are: 1) detect variant attack scenarios on different car models 2) does not require labeled data 3) works well even with a limited dataset. The effectiveness of the proposed model is evaluated on the survival dataset, and the experiment results show that the DAdAE model improves the overall f1 score significantly, compared to other unsupervised models.

Machine perception of emotions is integral to the development of human-centric Artificial Intelligence (AI) in sustainable industrial applications. Human expressions of emotions are not always direct. Word embeddings are mature techniques that can extract the semantics of such indirect expressions from text data. However, they are not primed to extract emotions. In this paper, we propose a novel approach that generates robust word embeddings for implicit and explicit expressions of emotion. This approach consists of two techniques, mask and rogue, we evaluate both techniques on two benchmark datasets for emotion classification. Our results confirm the effectiveness of the proposed approach in extracting emotions from diverse contexts. We have shared the emotion word embedding for public use.

Recent year has brought considerable advancements in Electric Vehicles (EVs) and associated infrastructures/communications. Intrusion Detection Systems (IDS) are widely deployed for anomaly detection in such critical infrastructures. This paper presents an Interpretable Anomaly Detection System (RX-ADS) for intrusion detection in CAN protocol communication in EVs. Contributions include: 1) Feature Extractor; 2) Anomaly Detection System; and 3) Explanation Generator for detected anomalies. The presented approach was tested on two benchmark CAN datasets: OTIDS and Car Hacking. The anomaly detection performance of RX-ADS was compared against the state-of-the-art approaches on these datasets: HIDS and GIDS. The RX-ADS approach showed comparable performance to the HIDS approach on OTIDS dataset and outperformed HIDS and GIDS approaches on Car Hacking dataset. Further, the proposed approach was able to generate explanations for detected abnormal behaviors arising from various intrusions. These explanations were later validated by information used by domain experts to detect anomalies. Other advantages of RX-ADS include: 1) the method can be trained on unlabeled data; 2) explanations help experts in understanding anomalies and root course analysis, and also help with AI model debugging and diagnostics, ultimately improving user trust in AI systems.

The industrial advancement of human society has been fundamentally driven by diverse ‘systems’ that facilitate ‘human interaction’ within physical, digital, virtual, social and artificial environments, and upon the hyper-connected layers of system-system interactions across these environments. The research and practice of Human System Interaction (HSI) has undergone exponential development due to the enhanced capabilities, increased efficiencies and decreased costs of digitalization. Primarily driven by its unique capacity for information persistence, digitalization is now leading us into a nexus of transition where HSI is being transformed by Artificial Intelligence (AI). AI has leveraged the data and information amassed by digitalization to learn, reason, predict, optimize and thereby augment both human-system interaction and system-system interaction, within and across all hyper-connected environments noted above. In this paper, we review this evolution of HSI and contribute towards its future directions by articulating the AI transformation strategy for this nexus of transition into a Human-AI-System Interaction. The paper begins with a review of HSI that focuses on developments in the past 15 years, followed by the AI transformation strategy which comprises of the primary configurations for Human-AI-System Interaction, the current capabilities of AI, a lifecycle approach for the design, development and deployment of an AI solution and the ethical implications of AI in HSI.

In critical infrastructures, timely detection of anomalies is essential to detect failures, avoid catastrophic damages, and improve resilience. Neural Network models are one of the state-of-the-art approaches used for anomaly detection. Among Neural Network architectures used these days, Autoencoders (AEs) have gained significant attention due to their advantages such as unsupervised learning, dimensionality reduction, non-linear feature extraction, the ease of integration with other neural network algorithms, and ease of use. Therefore, in this paper, we present: 1) anomaly detection and types of anomaly detection, 2) recent advancements in AEs typically used in anomaly detection, 3) AE-based Anomaly Detection (AE-AD) in selected critical infrastructures such as smart grids, intelligent transportation systems, and smart buildings, and 4) future research opportunities. We hope that this systematic survey of AE-based anomaly detection approaches will help the community prioritize research efforts to address pressing issues in critical infrastructures.

Clustering is an AI technique that has been successfully applied to the abundance of unlabelled real-world data for revealing hidden patterns and knowledge extraction. Deep Embedded Clustering (DEC) is a deep Autoencoder (AE) based model that learns feature representations and cluster assignments simultaneously. DEC learns the mapping from input data to a low-dimensional embedded space through joint optimization of feature transformation and clustering. Our previous work demonstrates how adding residual connections to deep AEs (RAEs) reduces the performance degradation of learned features when performing downstream classification on learned features. Further, it evidenced that RAE has improved unsupervised feature learning capability compared to AE. In this paper, we are evaluating the effect of residual connections in the context of Deep Embedded Clustering (DEC), which we refer to as RDEC. RDEC was compared against regular DEC. We considered various numbers of hidden layers and several bench-mark datasets: MNIST, Fashion MNIST, Reuters, and Human activity recognition. When increasing the depth of the neural network gradually, the presented RDEC showed up to 56% of less performance degradation compared to DEC. Further, the distribution of clustering accuracies showed that the presented RDEC outperforms DEC when comparing the accuracy variance and mean accuracy.

Despite the performance of state-of-the-art Artificial Intelligence (AI) systems, some sectors hesitate to adopt AI because of a lack of trust in these systems. This attitude is prevalent among high-risk areas, where there is a reluctance to remove humans entirely from the loop. In these scenarios, Augmentation provides a preferred alternative over complete Automation. Instead of replacing humans, AI Augmentation uses AI to improve and support human operations, creating an environment where humans work side by side with AI systems. In this paper, we discuss how AI Augmentation can provide a path for building Trustworthy AI. We exemplify this approach using Robot Teleoperation. We lay out design guidelines and motivations for the development of AI Augmentation for Robot Teleoperation. Finally, we discuss the design of a Robot Teleoperation testbed for the development of AI Augmentation systems.

Artificial Intelligence (AI) is influencing almost all areas of human life. Even though these AI-based systems frequently provide state-of-the-art performance, humans still hesitate to develop, deploy, and use AI systems. The main reason for this is the lack of trust in AI systems caused by the deficiency of transparency of existing AI systems. As a solution, “Trustworthy AI” research area merged with the goal of defining guidelines and frameworks for improving user trust in AI systems, allowing humans to use them without fear. While trust in AI is an active area of research, very little work exists where the focus is to build human trust to improve the interactions between human and AI systems. In this paper, we provide a concise survey on concepts of trustworthy AI. Further, we present trustworthy AI development guidelines for improving the user trust to enhance the interactions between AI systems and humans, that happen during the AI system life cycle.

Object categorization in images is fundamental to various industrial areas, such as automated visual inspection, fast image retrieval, and intelligent surveillance. Most existing methods treat visual features (e.g., scale-invariant feature transform) as content information of the objects, while regarding image tags as their contextual information. However, the image tags can hardly be acquired in completely unsupervised settings, especially when the image volume is too large to be marked. In this article, we propose a novel contextual multivariate information bottleneck (CMIB) method to conduct unsupervised image object categorization in multiple visual contexts. Unlike using manual contexts, the CMIB method first automatically generates a set of high-level basic clusterings by multiple global features, which are unprecedentedly defined as visual contexts since they can provide overall information about the target images. Then, the idea of the data compression procedure for object category discovery is proposed, in which the content and multiple visual contexts are maximally preserved through a “bottleneck.” Specifically, two Bayesian networks are initially built to characterize the relationship between data compression and information preservation. Finally, a novel sequential information-theoretic optimization is proposed to ensure the convergence of the CMIB objective function. Experimental results on seven real-world benchmark image datasets demonstrate that the CMIB method achieves better performance than the state-of-the-art baselines.

Deep Neural Networks (DNNs) have limited ability to explain their acquired knowledge or decision rationale. As a result, end-users perceive DNNs as black-boxes and are hesitant to fully adopt them in safety-critical applications. Therefore, developing explainable DNNs has become a prime interest in neural network research. This paper presents a methodology for linguistically explaining the knowledge a DNN classifier has acquired in training. The main objective is to help users understand what the DNN has learned about each class. The presented methodology is fuzzy logic based and involves end-users of the system in the explanation process, enabling users to customize the explanations to match their requirements. This paper presents the explanation methodology, metrics of explanation quality, validation steps, and a discussion of advantages and limitations. The explanation methodology was implemented on a benchmark classification problem. Experimental results demonstrated the method's capability to explain the DNN-knowledge and validated the explanations.

Deep brain stimulation (DBS) is well recognized as an effective treatment for symptoms of movement disorders such as Parkinson's disease (PD), Essential Tremor, and dystonia. The selection of the appropriate contact on the DBS lead for optimal clinical efficacy can be challenging, particularly when considering directional leads. Electroencephalograms (EEG) and electrocorticography has been utilized to better understand the pathophysiology of PD but a methodology to provide an objective biomarker of effective stimulation has yet to be developed. Using machine learning techniques for feature extraction and classification, we contrast high resolution EEG captured during DBS against its resting state counterpart with the DBS off. We demonstrate, using 16 patients under DBS treatment for movement disorders, EEG's informative capacity to detect both effective DBS and the region undergoing stimulation.

A viable solution for increasing fuel efficiency in vehicles is optimizing driver behavior. In our previous work, we proposed a data-driven Intelligent Driver System (IDS), which calculated an optimal driver behavior profile for a fixed route. During operation, the optimal behavior was prompted to the drivers to guide their behavior toward improving fuel efficiency. This system was proposed for fleet vehicles mainly because a small increase in fuel efficiency of fleet vehicles has a significant impact on the economy. The system was tested on a portion of the fleet's route (12km) and achieved 9-20% of fuel saving. One limitation of the IDS was that the prompted behavior profile was the same for all drivers. However, the approach of driving is significantly different from driver to driver. Therefore, it is important to capture those differences in the optimal behavior profile creation and prompting. This paper presents the first steps of a modified IDS that incorporates different approaches of drivers in optimal behavior profile creation. This work has three main components: 1) analyzing the capability of scaling our previously proposed IDS to the complete route of the fleet, 2) assessing the capability of identifying different types of driver behavior from data, and 3) proposing an IDS framework for integrating different driver behavior in optimizing driver behavior. Experimental results showed that the existing IDS was able to achieve 26-37% estimated fuel savings on the complete route. Conclusions of the paper are: 1)the existing IDS scaled to longer routes, and 2) It is possible to identify different driver behavior using data.

Crowdsourcing applications are proven to be a promising tool to gather valuable information, which can be used for a wide range of tasks, such as ensuring public safety. Traffic data collected using these applications have been used for efficient evacuation planning in large cities. In this paper, we propose to use regression-based machine learning methods to predict hourly taxi rides for a given location in a target day of week and month. The presented method can be used for the following purposes: 1) Predicting the number of taxi rides for a given location at a given time, 2) Identifying hot spots in a city, 3) Getting a rough count of the population density at a given location at a targeted hour, and 4) Planing evacuation routes for possible disasters. The presented approach has potential use for resource planning and evacuation in large cities. The Taxi and Limousine Commission (TLC) trip record data collected from 2017 to 2018 was used for this experiment. It was found that random forest regression can successfully predict hourly taxi drop-offs for a given taxi zone as well as for the entire city of New York.

Characterizing communications in smart-grid distributed control systems is fundamental for understanding the expected behavior and identify abnormal scenarios. In this paper, we present a stochastic data-driven approach to model the the communication network in smart-grid systems. Our approach uses Mixture Poisson distributions to model the packet communication between the network devices. The network is modeled using a directed graph, where each edge represents a Poisson distribution of the packets being transmitted. Parameters are learned using mini-batch Expectation Maximization in order to scale to large datasets. The advantages of the presented approach are 1) unsupervised data-driven discovery of representative communication patterns, 2) intuitive visualization of the expected behavior, 3) scalability to large datasets, and 4) coherent and interpretable model. Tests were conducted in a simulated SCADA microgrid distributed control system environment.

The papers in this special section explore the resilience in the energy markets. The transformation of energy infrastructure is making the resilience a fundamental topic for the definition of future configuration of the energy systems. This is due to the increased penetration of renewable energy sources (RES), the distributed energy management, the change in the power demand, the integration between energy and mobility sectors, the progressive liberalization of energy markets, the increased probability of extreme weather events, and the occurrence of errors in the interaction between humans and machines.

Artificial neural networks (ANNs) have been frequently used in industrial applications to model complex systems. However, using traditional ANNs for long-term planning tasks remains a challenge as they lack the capability to model uncertainty. Process noise and approximation errors cause ANN long-term estimations to deviate from the real behavior of the system. Unlike traditional ANNs, stochastic models provide a natural way to model uncertainty, providing estimations over a range of several possible outcomes. This paper introduces a stochastic modeling and planning approach using deep Bayesian neural networks (DBNNs). We use DBNNs to learn a stochastic model of the system dynamics. Planning is addressed as an open-loop trajectory optimization problem. We present two approaches for learning the dynamics: using single-step predictions and using multistep predictions. The advantages of the proposed methodology are as follows. First, accurate long-term estimations of the system state-trajectory probability distribution without the need for expert knowledge of the dynamics. Second, improved generalization and faster convergence rates in the trajectory optimization task when using multistep predictions to train the model. Third, viable for real-world applications since all expensive optimizations are executed offline while using a reasonable number of data samples. Testing is performed using challenging underactuated benchmark problems: the Cartpole and the Acrobot. The presented methodology successfully learns the swing-up maneuver using a relatively small number of iterations, with less than 125 sampled trajectories, and without any expert knowledge of the dynamics.

The deep self-organizing map (DSOM) was introduced to embed hierarchical feature abstraction capability to self-organizing maps (SOMs). This paper presents an extended version of the original DSOM algorithm (E-DSOM). E-DSOM enhances the DSOM in two ways-learning algorithm is modified to be completely unsupervised, and architecture is modified to learn features of different resolution in hidden layers. E-DSOM has three main advantages over the original DSOM: 1) improved classification accuracy; 2) improved generalization capability; and 3) need of fewer sequential layers (reduced training time). E-DSOM was tested on benchmark and real-world datasets and was compared against DSOM, SOM, sStacked autoencoder (AE), and stacked convolutional autoencoder (CAE). Experimental results showed that the E-DSOM outperformed DSOM with improvements of classification accuracy up to 15% while saving training time up to 19% on all datasets. Moreover, E-DSOM evidenced better generalization capability compared to the DSOM by showing superior performance on all datasets with induced noise. Further, E-DSOM showed comparable performance to the AE and the CAE while outperforming them on two datasets.

Anomaly detection in industrial processes is crucial for general process monitoring and process health assessment. Deep Neural Networks (DNNs) based anomaly detection has received increased attention in recent work. Albeit their high accuracy, the black-box nature of DNNs is a drawback in practical deployment. Especially in industrial anomaly detection systems, explanations of DNN detected anomalies are crucial. This paper presents a framework for DNN based anomaly detection which provides explanations of detected anomalies. The framework answers the following questions during online processing: 1) “why is it an anomaly?” and 2) “what is the confidence?” Further, the framework can be used offline to evaluate the “knowledge” of the trained DNN. The framework reduces the opaqueness of the DNN based anomaly detector and thus improves human operators' trust in the algorithm. This paper implements the first steps of the presented framework on the benchmark KDD-NSL dataset for Denial of Service (DoS) attack detection. Offline DNN explanations showed that the DNN was detecting DoS attacks based on features indicating destination of connection, frequency and amount of data transferred while showing an accuracy around 97%.

Data-driven models provide a powerful and flexible modeling framework for decision making and controls in industry. However, extracting knowledge from these models requires development of easily interpretable visualizations. In this paper, we present a data-driven methodology for modeling and visualization of relative equipment workload in a biomass feedstock preprocessing plant. The methodology is designed to serve in two main fronts: (1) knowledge discovery and data-mining from instrumentation data, (2) improving situational awareness during monitoring and control of the plant. We used Gaussian Processes to create a model of the expected current overload rate of for each of the electric motors involved in the plant. The expected number of overloads on each equipment was used to quantify and visualize the relative workload of the different components of the system. The visualization is presented in the form of an intuitive directed graph, whose properties (node size, position, colors) are driven by overload rates estimations.

Visual data mining facilitates the involvement of domain experts in the data mining processes. The effectiveness of visual data mining is especially dominant when paired with unsupervised methods due to the abundance of unlabeled data. Deep Self-Organizing Maps (DSOMs) are unsupervised learning architectures capable of high level feature abstraction. In this paper, we analyze the effectiveness of using DSOMs for visual data mining. DSOM's visual data mining capability was evaluated using the following visual data explorations methodologies: 1) U-Matrix, 2) hit maps and 3) data histograms. In comparison with traditional single layered SOM architectures, experimental results showed that DSOMs produced more accurate visual representations of the underlying data distributions. Therefore, DSOM is a viable method for generating easily understandable visual representations of high-dimensional complex datasets. These visual representations can be powerful tools in the real world, leading to better understanding of systems and thus enabling the design of better algorithms for control and monitoring.

Deep Neural Networks based intrusion detection systems (DNN-IDS) have proven to be effective. However, in domains like critical infrastructure security, user trust on the DNN-IDS is imperative and high accuracy isn't sufficient. The black-box nature of DNNs hinders transparency of the DNN-IDS, which is necessary for building trust. The main objective of this work is to improve user trust by improving transparency of the DNN-IDS by making it more communicative. This paper presents a methodology to generate offline and online feedback to the user on the decision making process of the DNN-IDS. Offline, the user is reported the input features that are most relevant in detecting each type of intrusion by the trained DNN-IDS. Online, for each detection, the user is reported the inputs features that contributed most to the detection. The presented method was implemented on the KDD-NSL dataset with a multi-layer perceptron (MLP) based DNN-IDS. Binary and multi-class classification was carried out on the dataset. Further, several DNN-IDS architectures with different depth were tested to study the factors that drive classification. It was observed that despite showing very similar accuracy results, the factors that drove the decisions were different across architectures. This evidences that the qualitative analysis that is enabled through reporting relevant input features is important for the user to make a more informed decision in choosing a DNN-IDS. This online and offline feedback leads to improving the transparency of the DNN-IDS and helps build trust prior to and during deployment.

Despite the growing popularity of modern machine learning techniques (e.g, Deep Neural Networks) in cyber-security applications, most of these models are perceived as a black-box for the user. Adversarial machine learning offers an approach to increase our understanding of these models. In this paper we present an approach to generate explanations for incorrect classifications made by data-driven Intrusion Detection Systems (IDSs) An adversarial approach is used to find the minimum modifications (of the input features) required to correctly classify a given set of misclassified samples. The magnitude of such modifications is used to visualize the most relevant features that explain the reason for the misclassification. The presented methodology generated satisfactory explanations that describe the reasoning behind the mis-classifications, with descriptions that match expert knowledge. The advantages of the presented methodology are: 1) applicable to any classifier with defined gradients. 2) does not require any modification of the classifier model. 3) can be extended to perform further diagnosis (e.g. vulnerability assessment) and gain further understanding of the system. Experimental evaluation was conducted on the NSL-KDD99 benchmark dataset using Linear and Multilayer perceptron classifiers. The results are shown using intuitive visualizations in order to improve the interpretability of the results.

Cyber-Physical Systems (CPSs)have become ubiquitous in recent years and has become the core of modern critical infrastructure and industrial applications. Therefore, ensuring security is a prime concern. Due to the success of Deep Learning (DL)in a multitude of domains, development of DL based CPS security applications have received increased interest in the past few years. Developing generalized models is critical since the models have to perform well under threats that they havent trained on. However, despite the broad body of work on using DL for ensuring the security of CPSs, to our best knowledge very little work exists where the focus is on the generalization capabilities of these DL applications. In this paper, we intend to provide a concise survey of the regularization methods for DL algorithms used in security-related applications in CPSs and thus could be used to improve the generalization capability of DL based cyber-physical system based security applications. Further, we provide a brief insight into the current challenges and future directions as well.

Epileptic seizure source identification involves neurologists combing through a substantial amount of data manually, which sometimes takes weeks per patient. This paper presents a methodology for minimizing the amount of data a neurologist has to analyze to identify the seizure focus. The method keeps the neurologist as the final decision maker and aids in the decision making process. It has to be noted that the primary focus of the work was not improving the accuracy of interictal spike detection but reduction of the volume of data. The presented methodology is based on Artificial Neural Networks (ANN) and is implemented on EEG data collected on 5 patients using a dense array EEG reader. As a baseline, a simple template matching was implemented on the same dataset. Experimental results showed that the ANN based methodology was able to reduce the dataset by 98%, a significant improvement on the template matching method.

Welcome to HSI2017, the 10th International Conference on Human System Interactions in 2017 was held at the University of Ulsan in Ulsan, Republic of Korea. The University of Ulsan have organized the conference and the conference is technically co-sponsored by IEEE Industrial Electronics Society. HSI conference series has been one of the most important academic meetings in the field of interactions between human and systems. Until now the HSI conference series have been held in Krakow (Poland) 2008, Catania (Italy) 2009, Rzeszow (Poland) 2010, Yokohama (Japan) 2011, Perth (Australia) 2012, Gdansk (Poland) 2013, Lisbon (Portugal) 2014, Warsaw (Poland) 2015, and Portsmouth (United Kingdom) 2016.

Control systems monitor and command other devices, systems, and software within an infrastructure. Typically, control systems employ human-in-the-loop control for critical decision making and response. These end-users require easy access to accurate, actionable and relevant data to ensure quick and effective decision making. This work presents a framework for creating dynamic visual interfaces for improved situational awareness. The proposed framework determines the relevance of available information pieces and then applies the derived relevance scores to a visualization so that the most relevant and important information are emphasized to the end-users. In the presented work, a priori expert knowledge is encoded in the system through the use of Fuzzy Logic (FL) and the resulting FL inference system assigns scores to information pieces based on system state information and user defined relevance. These scores can then be used to organize and display the relevant data given the current situation and end-user roles. The proposed FL based scoring system was implemented on a real world control system dataset and we demonstrate how the information visualization is dynamically adapted to improve situational awareness. Further, we discuss potential methods the relevance scores can be incorporated into real world visualizations to increase the situational awareness in control systems.

Artificial neural networks and deep learning methodologies have had growing interest across industry domains, including IoT and mobile systems. However, in low-power applications, resource limitations and operating environment restrictions make implementations difficult. This survey examines efforts that target the data and compute challenges of implementing energy efficient, low cost, and accurate neural network models. Approaches come in many forms, with solutions ranging from software optimization to hardware reorganization. We examine three avenues of approach - binary neural networks, application specific circuit designs, and neuromorphic computing. For each methodology, we summarize progress, use-cases, and inherent challenges.

Disk arrays provide data storage by combining sets of disks into one or more virtual disks (VDs) utilizing specialized control algorithms. A VD is a partition of such combined storage capacity perceived by the user as a physical disk. VD cloning is a data protection technique that copies the data from one VD to another. A typical problem with VD cloning is that it delays user reads and writes, i.e., increases response times. This paper presents VD- Interval Type 2 (IT2), an effective IT2 fuzzy logic control (IT2 FLC) approach to VD cloning that dramatically reduces the response time delay. The VD-IT2 cloning scheme is capable of balancing two adversely affecting processes: VD cloning and user read/write response times. This IT2 FLC-based approach regulates throughput of VD cloning with regards to the user request activity. The VD-IT2 solves the increased delay in user reads and write during cloning by adjusting the time interval between cloning requests. The contribution of this paper is two-fold. First, a novel formula for planning of data backups is presented. The data backup formula predicts the fraction of replicated data blocks in a combined (clone and snap) replication. Second, a novel IT2 FLC scheme reduces response time of user reads/writes in half. The VD-IT2 was tested on an Itanium workstation with 120 disks and proved 50% reduction of user latencies within a short period of time (high response time of 60 ms was reduced to half, within 30 s only), on an average basis.

Machine control using electroencephalography (EEG) based brain computer interfaces (BCI) has been extensively researched in the past decade. However, research is often based on event bound methods such as motor imagery. Despite being useful in medical applications, even bound methods limit users' operational capability while performing BCI control. To alleviate the said limitation, we explore a robot control framework based on abstract thought. Abstract thought in this context is defined as conscious mental tasks that are not bound with any particular event or bodily movement. This paper presents an initial step in the framework, which is a methodology for optimal feature selection for abstract thought EEG data classification. The presented method contains 2 steps: 1) generational Genetic Algorithm (GA) based feature selection, and, 2) EEG data classification using selected features. The presented method was implemented on an EEG dataset acquired from a consumer grade EEG device. Abstract thought EEG data were collected for three actions pertaining to robot control; 1) “rest”, 2) “move forward”, and, 3) “turn left”. The presented method was compared to EEG classification without any feature selection. Experimental results showed that the presented method outperformed the method without feature selection for all the tested classifiers with a 10% or higher improvement in classification accuracy.

Epilepsy is a neurological disease that causes seizures in its victims that can lead to physical injury or even death in some circumstances. It is caused by excessive, synchronous abnormal firing of neurons in the brain. This chronic disease has no known cure and affects millions of people worldwide but can be managed through various methods. The successful treatment is dependent upon correct identification of the origin of the seizures within a brain. One major challenge for doctors is the analysis of the immense amount of data collected by electroencephalogram (EEG) devices. In order to identify a region of the brain that causes epileptic seizures, millions of samples must be analyzed manually by a trained eye to find interictal spikes that emanate from the afflicted region of the brain. This paper presents a method for automatic interictal spike detection while minimizing false positives. In this way, it eliminates the lengthy, manual process currently used by doctors. Analyzing real world data, the presented Neural Network Epileptic Spike Detector (NNESD) showed a PPV of 72.67% and sensitivity of 82.68% on average over 300 trained networks on a single channel of EEG.

The efficient use of energy is one method of potentially reducing the amount of fossil fuels that are used world-wide to generate electricity. Nuclear energy presents a promising sustainable energy source for the future that generates few to no greenhouse gases. While it creates energy that can be used for electricity it also has many byproducts that are treated as waste or not utilized to their full potential. Waste heat is a major product of nuclear energy that should be harnessed and applied to other processes. These include water desalination/water purification hydrogen production, use in industrial chemical operations, or electricity production. Nuclear energy is most often base load following power and very inflexible. One way to address this is to use Hybrid Energy Systems (HES). By doing this the system can be load following and hence more efficient and sustainable. Therefore, the goal of this project is to create a system that mimics the waste heat from a reactor, demonstrate how to utilize that heat, and show that when energy demand is low the reactor does not need to reduce power; the energy can be directed elsewhere to create goods. This paper presents the details of the Energy Conversion Loop (ECL) that was developed to act as a test bed for experimentation of the aforementioned processes and test the possibility of nuclear HES. Further, the paper presents a proposed intelligent control system that can adapt to the system requirements for the ECL. In addition, the system has great potential for education on critical infrastructure protection and testing of future control logic systems and mechanical systems.

Ensuring sustainability demands more efficient energy management with minimized energy wastage. Therefore, the power grid of the future should provide an unprecedented level of flexibility in energy management. To that end, intelligent decision making requires accurate predictions of future energy demand/load, both at aggregate and individual site level. Thus, energy load forecasting have received increased attention in the recent past. However, it has proven to be a difficult problem. This paper presents a novel energy load forecasting methodology based on Deep Neural Networks, specifically, Long Short Term Memory (LSTM) algorithms. The presented work investigates two LSTM based architectures: 1) standard LSTM and 2) LSTM-based Sequence to Sequence (S2S) architecture. Both methods were implemented on a benchmark data set of electricity consumption data from one residential customer. Both architectures were trained and tested on one hour and one-minute time-step resolution datasets. Experimental results showed that the standard LSTM failed at one-minute resolution data while performing well in one-hour resolution data. It was shown that S2S architecture performed well on both datasets. Further, it was shown that the presented methods produced comparable results with the other deep learning methods for energy forecasting in literature.

Dependency of the transport sector on fossil fuels is encouraging a significant amount of research in to improving fuel efficiency in vehicles. Three primary techniques are identified for vehicle fuel efficiency improvement: 1) vehicle technology improvements such as drivetrain improvements, 2) traffic infrastructure improvements such as traffic flow management and route selection, and 3) driver behavior changes such as acceleration and deceleration profiles. Out of the 3 techniques, driver behavior changing has the least implementation cost and is able to provide immediate results. Thus, this paper presents a fuel efficient driving behavior identification and feedback architecture that is specific to fleet vehicles. The presented method utilizes historical data from fleet drivers on specific routes and generates fuel optimal velocity profiles that do not affect travel time. The identified velocity profile is the prompted to the driver via a low-cost plug-and-play style un-obstructive display. The display uses an intuitive and easily understandable visualization to prompt drivers on fuel efficient velocity. The presented architecture was tested on the Idaho National Laboratory (INL) bus fleet in real-world driving conditions and was shown to be able to increase the fuel economy by 9% and 20% in two different driving scenarios.

Heating, Ventilation and Air Conditioning (HVAC) system is largest energy consumer in buildings. Worldwide, buildings consume 20% of the total energy production. Therefore, increasing efficiency of the HVAC system will result in significant financial savings. As one solution, Thermal Energy Storage (TES) tanks are being utilized with buildings to store excess energy to be reused later. An optimal control strategy is crucial for optimal usage. Therefore, this paper presents a novel control framework based on Artificial Neural Networks (ANN) for optimally controlling a TES for achieving increased savings. The presented ANN controller utilizes 3 main inputs: 1) current TES energy availability, 2) predicted building power requirement, and 3) predicted utility load/price. In addition to the design details of the control framework, this paper presents implementation details of the ANN controller. Further, experiments on several test cases were carried out and the paper presents the experimental setup and obtained results for each test case. Performance of the presented ANN control framework was compared against a classical proportional derivative (PD) controller. It was observed that the presented framework resulted in better cost savings than the classical controller consistently for all the experimental test cases.

Buildings are known to be significant energy consumers throughout the world. Thus, improving the energy efficiency of buildings is a key research goal. However, maintaining occupant comfort while improving energy efficiency in buildings requires close monitoring of the building environment and immediate control actions taken when sub-optimal behavior is identified. Such monitoring requires high frequency data from sensors. Therefore, increasing the data collection rate or the temporal resolution of sensors can lead to improved building control and state-awareness. This paper presents an on-line learning, data-fusion based methodology that uses Artificial Neural Networks (ANNs) to increase temporal resolution of building sensor data. The presented method utilizes sensor information from different sensors in the building to predict higher temporal resolution data of specific sensors. Furthermore, the presented method is capable of learning changing building behavior for improved long-term accuracy. The presented method was applied to a real-world building dataset and was shown to be able to predict high temporal resolution data with a higher accuracy compared to classical methods. Furthermore, the on-line learning was shown to increase the prediction accuracy in long-term operation.

Fast and effective monitoring following airborne releases of toxic substances is critical to mitigate risks to threatened population areas. Electrically powered systems in industrial settings require monitoring of emitted electromagnetic fields to determine the status of the equipment and ensure their safe operation. In situations such as these, wireless sensor nodes (WSNs) at fixed predetermined locations provide monitoring to ensure safety. A challenging algorithmic problem is determining the locations to place these WSNs while meeting several criteria: (1) to provide complete coverage of the domain; (2) to create a topology with problem-dependent node densities; and (3) to minimize the number of WSNs. This paper presents a novel approach, advancing front mesh generation with constrained Delaunay triangulation and smoothing (AFECETS) that addresses these criteria. A unique aspect of AFECETS is the ability to determine WSN locations for areas of high interest (hospitals, schools, and high population density areas) that require higher density of nodes for monitoring environmental conditions, a feature that is difficult to find in other research work. The AFECETS algorithm was tested on several arbitrary shaped domains. AFECETS simulation results show that the algorithm provides significant reduction in the number of nodes, in some cases over 40%, compared with an advancing front mesh generation algorithm; maintains and improves optimal spacing between nodes; and produces simulation run times suitable for real-time applications.

Generalized solutions to complex problems often suffer from being overly complicated. The main contribution of this paper is to describe an architecture that allows for greater problem generalization without the traditional corresponding increase in complexity. The architecture extends traditional fuzzy logic and is called Fuzzy Contexts or Fuzzy Logic Type-C. Fuzzy logic permits partial membership and values can belong to multiple fuzzy sets. By breaking down a problem space into smaller contexts and allowing algorithms themselves to have relaxed memberships in those contexts, a Type-C solution can support multiple solutions to complex problems. This paper describes how problem spaces may be decomposed into smaller, more easily solvable components and fuzzified together under a Type-C hierarchy. Test results with a simulated robotic navigation system demonstrates how a Type-C implementation is able to improve upon a generalized fuzzy controller.

Fuzzy Logic Systems (FLS) are a well documented proven method for various applications such as control classification and data mining. The major advantage of FLS is the use of human interpretable linguistic terms and rules. In order to capture the uncertainty inherent to linguistic terms, Fuzzy Membership Functions (MF) are used. Therefore, membership functions are essential for improving the understandability of fuzzy systems. Optimizing FLS for improved accuracy in terms of classification or control can reduce the understandability of fuzzy MFs. Expert knowledge can be used to derive MFs, but it has been shown that this might not be optimal, and acquiring expert knowledge is not trivial. Therefore, this paper presents a data driven method using statistical methods to generate membership functions that describe the data while maintaining the understandability. The presented method calculates key points such as membership function centers, intersections and slopes using data driven statistical methods. Furthermore, the presented method utilizes several understandability metrics to adjust the generated MFs. The presented method was tested on several benchmark datasets and a real-world dataset and was shown to be able to generate MFs that describe the dataset, while maintaining high levels of understandability.

Brain Computer Interfaces (BCI) have gained significant interest over the last decade as viable means of human machine interaction. Although many methods exist to measure brain activity in theory, Electroencephalography (EEG) is the most used method due to the cost efficiency and ease of use. However, thought pattern based control using EEG signals is difficult due two main reasons; 1) EEG signals are highly noisy and contain many outliers, 2) EEG signals are high dimensional. Therefore the contribution of this paper is a novel methodology for recognizing thought patterns based on Self Organizing Maps (SOM). The presented thought recognition methodology is a three step process which utilizes SOM for unsupervised clustering of pre-processed EEG data and feed-forward Artificial Neural Networks (ANN) for classification. The presented method was tested on 5 different users for identifying two thought patterns; “move forward” and “rest”. EEG Data acquisition was carried out using the Emotiv EPOC headset which is a low cost, commercial-off-the-shelf, noninvasive EEG signal measurement device. The presented method was compared with classification of EEG data using ANN alone. The experimental results for the 5 users chosen showed an improvement of 8% over ANN based classification.

Fuzzy logic provides a natural and precise way for humans to define and interact with systems. Optimizing a fuzzy inference system, however, presents some special challenges for the developer because of the imprecision that is inherent to fuzzy sets. This paper expands upon an earlier development of a fuzzy framework, adding components for dynamic self-optimization. What makes this approach unique is the use of relational database as a computational engine for the memetic algorithm and fitness function. The new architecture combines the power of fuzzy logic with the special properties of a relational database to create an efficient, flexible and self-optimizing combination. Database objects provide the fitness function, population sampling, gene crossover and mutation components allowing for superior batch processing and data mining potential. Results show the framework is able to improve the performance of a working configuration as well as fix a non-working configuration.

Robot remote teleoperation enables users to perform complex tasks in hostile or inaccessible environments, without physical presence. However, minimizing collisions with obstacles while maintaining accuracy and speed of task is important. While visual and auditory inputs to the user aid in accurate control, to achieve the required speed and accuracy, tactile and kinesthetic force-feedback information can be used. This paper presents a dynamic real-time fuzzy logic based force-feedback control for obstacle avoidance in a remotely operated robot manipulator. The presented method utilizes absolute position of the robot manipulator to calculate the distance vector to known obstacles. A fuzzy controller utilizes the distance vectors and the velocities of the components in the manipulator to generate force feedback in each axis. Furthermore, the paper presents an interactive graphical user interface that enables users to add or remove obstacles in the environment dynamically. The presented method was implemented on a simple 3-DOF robot manipulator. The presented method was compared to a situation without force feedback. Test results show significantly improved speed and consistency in completing a task when the presented force feedback method is used.

With escalating fuel prices, limited fossil fuel reserves and increasing carbon emissions, significant efforts are being made to decrease fuel consumption in vehicles. While, drivetrain improvements play a major role in improving fuel economy, it has been identified that fuel efficient driving behavior is a viable method for increased fuel efficiency. Thus, if the optimal fuel efficient behavior can be identified, it can be used to increase the fuel efficiency of drivers. However, once the optimal fuel efficient behavior is identified, it has to be presented to the driver, while the vehicle is being driven. Thus, this method of information representation has to be un-obstructive and easy to comprehend. This paper presents a low cost framework and a hardware setup for prompting drivers on fuel efficient behavior. The presented framework includes an information rich, intuitive un-obstructive visualization. The presented method was implemented using low cost, commercial-off-the-shelf hardware and tested on a sample of buses selected from the Idaho National Laboratory (INL) bus fleet. Different types of visual cues were and evaluated by professional drivers for obstructiveness, interpretability and intuitiveness.

Advanced remote teleoperation of robot manipulators enable complex tasks to be performed in hostile or inaccessible environments, without the physical presence of a human. For increased effectiveness of teleoperation, maintaining accuracy and speed of task while minimizing collisions is important. Visual and auditory inputs to the user aid in accurate control. However, to further increase the speed and accuracy, tactile and kinesthetic force-feedback information can be used. One of the most common methods of force-feedback generation is the virtual force field based method. However, in complex environments where increased accuracy is required, static force field based methods are insufficient. This paper presents a dynamically varying, virtual force field based force-feedback generation method for obstacle avoidance in remotely operated robot manipulators. The presented method utilizes a fuzzy logic model to dynamically vary a virtual force field surrounding the manipulator in real-time. The fuzzy controller utilizes the distance vectors to obstacles and the velocity vectors of the manipulator components to generate the force field in each axis. The generated force field is then used to calculate the final force-feedback that is sent to the user. The presented method was implemented on a simple 3-DOF robot manipulator, and compared to a typical static force field based force-feedback generation method. Test results show that the task completion time is significantly improved without significant loss of accuracy in certain tasks when the presented force-feedback method is used.

As critical and sensitive systems increasingly rely on complex software systems, identifying software vulnerabilities is becoming increasingly important. It has been suggested in previous work that some bugs are only identified as vulnerabilities long after the bug has been made public. These bugs are known as Hidden Impact Bugs (HIBs). This paper presents a hidden impact bug identification methodology by means of text mining bug databases. The presented methodology utilizes the textual description of the bug report for extracting textual information. The text mining process extracts syntactical information of the bug reports and compresses the information for easier manipulation. The compressed information is then utilized to generate a feature vector that is presented to a classifier. The proposed methodology was tested on Linux vulnerabilities that were discovered in the time period from 2006 to 2011. Three different classifiers were tested and 28% to 88% of the hidden impact bugs were identified correctly by using the textual information from the bug descriptions alone. Further analysis of the Bayesian detection rate showed the applicability of the presented method according to the requirements of a development team.

Resiliency and improved state-awareness of modern critical infrastructures, such as energy production and industrial systems, is becoming increasingly important. As control systems become increasingly complex, the number of inputs and outputs increase. Therefore, in order to maintain sufficient levels of state-awareness, a robust system state monitoring must be implemented that correctly identifies system behavior even when one or more sensors are faulty. Furthermore, as intelligent cyber adversaries become more capable, incorrect values may be fed to the operators. To address these needs, this paper proposes a fuzzy-neural data fusion engine (FN-DFE) for resilient state-awareness of control systems. The designed FN-DFE is composed of a three-layered system consisting of: 1) traditional threshold based alarms; 2) anomalous behavior detector using self-organizing fuzzy logic system; and 3) artificial neural network-based system modeling and prediction. The improved control system state-awareness is achieved via fusing input data from multiple sources and combining them into robust anomaly indicators. In addition, the neural network-based signal predictions are used to augment the resiliency of the system and provide coherent state-awareness despite temporary unavailability of sensory data. The proposed system was integrated and tested with a model of the Idaho National Laboratory's hybrid energy system facility known as HYTEST. Experiment results demonstrate that the proposed FN-DFE provides timely plant performance monitoring and anomaly detection capabilities. It was shown that the system is capable of identifying intrusive behavior significantly earlier than conventional threshold-based alarm systems.

A challenge facing industrial control network administrators is protecting the typically large number of connected assets for which they are responsible. These cyber devices may be tightly coupled with the physical processes they control and human induced failures risk dire real-world consequences. Dynamic virtual honeypots are effective tools for observing and attracting network intruder activity. This paper presents a design and implementation for self-configuring honeypots that passively examine control system network traffic and actively adapt to the observed environment. In contrast to prior work in the field, six tools were analyzed for suitability of network entity information gathering. Ettercap, an established network security tool not commonly used in this capacity, outperformed the other tools and was chosen for implementation. Utilizing Ettercap XML output, a novel four-step algorithm was developed for autonomous creation and update of a Honeyd configuration. This algorithm was tested on an existing small campus grid and sensor network by execution of a collaborative usage scenario. Automatically created virtual hosts were deployed in concert with an anomaly behavior (AB) system in an attack scenario. Virtual hosts were automatically configured with unique emulated network stack behaviors for 92% of the targeted devices. The AB system alerted on 100% of the monitored emulated devices.

The proliferation of digital devices in a networked industrial ecosystem, along with an exponential growth in complexity and scope, has resulted in elevated security concerns and management complexity issues. This paper describes a novel architecture utilizing concepts of autonomic computing and a simple object access protocol (SOAP)-based interface to metadata access points (IF-MAP) external communication layer to create a network security sensor. This approach simplifies integration of legacy software and supports a secure, scalable, and self-managed framework. The contribution of this paper is twofold: 1) A flexible two-level communication layer based on autonomic computing and service oriented architecture is detailed and 2) three complementary modules that dynamically reconfigure in response to a changing environment are presented. One module utilizes clustering and fuzzy logic to monitor traffic for abnormal behavior. Another module passively monitors network traffic and deploys deceptive virtual network hosts. These components of the sensor system were implemented in C++ and PERL and utilize a common internal D-Bus communication mechanism. A proof of concept prototype was deployed on a mixed-use test network showing the possible real-world applicability. In testing, 45 of the 46 network attached devices were recognized and 10 of the 12 emulated devices were created with specific operating system and port configurations. In addition, the anomaly detection algorithm achieved a 99.9% recognition rate. All output from the modules were correctly distributed using the common communication structure.

Building Energy Management Systems (BEMSs) are essential components of modern buildings that are responsible for minimizing energy consumption while maintaining occupant comfort. However, since indoor environment is dependent on many uncertain criteria, performance of BEMS can be suboptimal at times. Unfortunately, complexity of BEMSs, large amount of data, and interrelations between data can make identifying these suboptimal behaviors difficult. This paper proposes a novel Fuzzy Anomaly Detection and Linguistic Description (Fuzzy-ADLD)-based method for improving the understandability of BEMS behavior for improved state-awareness. The presented method is composed of two main parts: 1) detection of anomalous BEMS behavior; and 2) linguistic representation of BEMS behavior. The first part utilizes modified nearest neighbor clustering algorithm and fuzzy logic rule extraction technique to build a model of normal BEMS behavior. The second part of the presented method computes the most relevant linguistic description of the identified anomalies. The presented Fuzzy-ADLD method was applied to real-world BEMS system and compared against a traditional alarm-based BEMS. Six different scenarios were tested, and the presented Fuzzy-ADLD method identified anomalous behavior either as fast as or faster (an hour or more) than the alarm based BEMS. Furthermore, the Fuzzy-ADLD method identified cases that were missed by the alarm-based system, thus demonstrating potential for increased state-awareness of abnormal building behavior.

Selecting the optimal dimensions for various knowledge extraction applications is an essential component of data mining. Dimensionality selection techniques are utilized in classification applications to increase the classification accuracy and reduce the computational complexity. In text classification, where the dimensionality of the dataset is extremely high, dimensionality selection is even more important. This paper presents a novel, genetic algorithm based methodology, for dimensionality selection in text mining applications that utilizes information gain. The presented methodology uses information gain of each dimension to change the mutation probability of chromosomes dynamically. Since the information gain is calculated a priori, the computational complexity is not affected. The presented method was tested on a specific text classification problem and compared with conventional genetic algorithm based dimensionality selection. The results show an improvement of 3% in the true positives and 1.6% in the true negatives over conventional dimensionality selection methods.

Fuzzy logic, Type-1 and Type-2, is well suited for human systems interactions because they provides a natural way of implementing linguistic terms from human experts. Existing fuzzy frameworks, however, provide limited support for Type-2. They also tend to be fairly complicated and/or have limited portability. This paper introduces a fuzzy framework for building a Type-1Type-2 fuzzy controller. A “wizard” application and modeling language are supported to provide an easy-to-use interface for creating a fuzzy inference system. The benefits of this framework are: (1) Increased understanding of fuzzy systems implementation via easy-to-use visual tools; (2) Reduced development time; (3) A standardized and portable codebase; (4) Easy configuration via XML; (5) Support for both Type-1 and Type-2 fuzzy sets and rules. The framework is tested and solves a maze problem using both Type-1 and Type-2 implementations.

Brain Computer Interfaces (BC!) are becoming increasingly studied as methods for users to interact with computers because recent technological developments have lead to low priced, high precision BCI devices that are aimed at the mass market. This paper investigates the ability for using such a device in real world applications as well as limitations of such applications. The device tested in this paper is called the Emotiv EPOC headset, which is an electroencephalograph (EEG) measuring device and enables the measuring of brain activity using 14 strategically placed sensors. This paper presents: 1) a BCI framework driven completely by thought patterns, aimed at real world applications 2) a quantitative analysis of the performance of the implemented system. The Emotiv EPOC headset based BCI framework presented in this paper was tested on a problem of controlling a simple differential wheeled robot by identifying four thought patterns in the user: “neutral”, “move forward”, “turn left”, and “turn right”. The developed approach was tested on 6 individuals and the results show that while BCI control of a mobile robot is possible, precise movement required to guide a robot along a set path is difficult with the current setup. Furthermore, intense concentration is required from users to control the robot accurately.

This is the first of a two-part investigation of centralized and decentralized approaches for determining the optimal configuration of a sensor network. In this first part, we present a centralized approach for the generation of mesh (wireless sensor) network configurations that provide complete sensing coverage and communication connectivity of a domain. A challenging problem in deploying wireless sensor networks is maximizing coverage in irregular shaped polygonal areas while maintaining a high degree of node connectivity. The novelties presented in this paper are: 1) a centralized mesh simplification technique, the Iterative Node Removal with Constrained Delaunay Triangulation and Smoothing (INRCDTS) algorithm, and 2) a centralized mesh generation approach with INRCDTS that may be used for any nonintersecting closed polygonal area. Additionally, we provide a comparison of two centralized mesh generation techniques. The INRCDTS was built and tested as an enhancement of two traditional mesh generation techniques: advancing front technique and Matlab partial differential equation toolbox. The INCRCDTS introduces the ability to tune the generated mesh configuration to the number of nodes and nodal spacing. The INRCDTS enhancement has proven to increase the uniformity of the mesh in an irregular shaped polygonal area relative to advancing front and MATLAB partial differential equation algorithms by 23% and 41%, respectively.

This is the second of a two-part investigation of the generation of wireless sensor network (WSN) configurations that: 1) maximize coverage of irregular shaped polygonal areas and 2) maintain a high degree of node connectivity. The first-part of the investigation presented centralized algorithms for the generation of mesh (wireless sensor) network configurations that maximize coverage and connectivity. In this second part, we present a decentralized and distributed approach using an Extended Virtual Spring Mesh (EVSM)-Adaptive Coverage Algorithm and Protocol (ACAP) algorithm. The EVSM-ACAP algorithm represents an extension of EVSM algorithm with the newly developed ACAP. ACAP provides adaptive coverage and configuration of the mesh network by dynamically adjusting the sensing range of sensor nodes. EVSM-ACAP is compared to centralized mesh generation algorithms (described in the part one of the investigation), as well as other decentralized algorithms from artificial physics, for the control of large numbers of physical agents in sensor networks. EVSM-ACAP is shown to produce a sensor network deployment with an average sensor spacing within 1.6% of the desired spacing, versus 5.75% for the best centralized algorithmic approach. To the best of our knowledge, this is the first time that these centralized mesh network configuration algorithms have been contrasted with the scalable, robust, decentralized algorithms of artificial physics and EVSM.

A mobile ad hoc network is a collection of independent nodes that communicate wirelessly with one another. This paper investigates nodes that are swarm robots with communications and sensing capabilities. Each robot in the swarm may operate in a distributed and decentralized manner to achieve some goal. This paper presents a novel approach to dynamically adapting control parameters to achieve mesh configuration stability. The presented approach to robot interaction is based on spring force laws (attraction and repulsion laws) to create near-optimal mesh like configurations. In prior work, we presented the extended virtual spring mesh (EVSM) algorithm for the dispersion of robot swarms. This paper extends the EVSM framework by providing the first known study on the effects of adaptive versus static control parameters on robot swarm stability. The EVSM algorithm provides the following novelties: 1) improved performance with adaptive control parameters and 2) accelerated convergence with high formation effectiveness. Simulation results show that 120 robots reach convergence using adaptive control parameters more than twice as fast as with static control parameters in a multiple obstacle environment.

Type-2 Fuzzy Logic Systems (T2 FLSs) have been commonly attributed with the capability to model various sources of data uncertainties. The input uncertainties of an FLS were modeled using T2 Fuzzy Sets (FSs) and the type-reduced centroid of the output FS was interpreted as a measure of uncertainty associated with the terminal real-valued output. However, the accuracy of this input-output uncertainty modeling has been rarely studied. It is well established that T2 FSs can be understood as a composition of a large number of embedded T1 FSs and thus model the uncertainty of selecting a specific T1 FSs. However, whether the same can be achieved with T2 FLSs can be considered an open question. This paper contributes by presenting a study of the input-output uncertainty modeling capability of Interval T2 (IT2) FLSs. First, the Monte Carlo simulation technique is used to simulate linguistic uncertainties and to compute the aggregated output result. This simulation is then compared to the output bounds provided by the interval centroid computed with IT2 FLS. It is demonstrated that the interval output of the IT2 FLS overestimates the output uncertainty range when compared to the results of the Monte Carlo simulation. To further understand this problem the concept of Equivalent Type-1 FSs is used. Finally, a detailed example is presented to demonstrate why the IT2 fuzzy inference process overestimates the output uncertainty.

General Type-2 Fuzzy Sets (GT2 FSs) have been originally proposed to allow for modeling uncertainty associated with the membership grades of Type-1 (T1) FSs. However, because of the computational complexity associated with the processing of GT2 FSs, only their constrained version, the Interval T2 (IT2) FSs, have been widely used. While IT2 FSs allow for fast processing, they lack the expressive power of GT2 FSs when modeling various kinds of uncertainties. In order to combine the best of both types, this paper proposes a novel class of T2 FSs - the Shadowed Type-2 (ST2) FSs. The ST2 FS is a T2 FS with secondary membership functions represented as Shadowed Sets (SSs). Shadowed sets, originally proposed by Pedrycz, are directly induced by the T1 fuzzy membership functions and they are designed to conserve the amount of uncertainty in the original T1 FS. In a similar manner, an ST2 FS is directly induced by a GT2 FS via transforming all the T1 fuzzy secondary membership functions into Shadowed Sets. The resulting ST2 FSs can thus better capture the uncertainty in the original GT2 FSs when compared to the constrained IT2 FSs. Additionally, ST2 FSs offer very efficient computational framework since the secondary membership grades can only attain three values of 0, 1, or completely uncertain (shadowed) grade of [0,1]. This paper introduces the representation, the elementary set-theoretic operations and several methods for type-reduction and defuzzification of ST2 FSs. The modeling capability of ST2 SS was demonstrated on several examples.

Reducing the fuel consumption of road vehicles has the potential to decrease environmental impact of transportation as well as achieve significant economical benefits. This paper proposes a novel methodology for improving the fuel economy of vehicle fleets via learning fuel-efficient driving behaviors. Vehicle fleets composed of large number of heavy vehicles routinely perform runs with different drivers over a set of fixed routes. While all drivers might achieve on-time and safe driving performance their actual driving behaviors and the subsequent fuel economy can vary substantially. The proposed Intelligent Driver System (IDS) utilizes vehicle performance data combined with GPS information on fixed routes to incrementally build a model of the historically most fuel efficient driving behavior. During driving, the calculated optimal velocity for specific location is compared to the current vehicle state and a fuzzy logic PD controller is used to compute the optimal control action. The control action can be projected to the drivers via a specialized HMI or used directly as a predictive cruise control to achieve overall fuel economy improvements. The method has been validated on a simulated heavy vehicle model, showing potential for substantial fuel economy improvements.

Identifying software vulnerabilities is becoming more important as critical and sensitive systems increasingly rely on complex software systems. It has been suggested in previous work that some bugs are only identified as vulnerabilities long after the bug has been made public. These vulnerabilities are known as hidden impact vulnerabilities. This paper discusses existing bug data mining classifiers and present an analysis of vulnerability databases showing the necessity to mine common publicly available bug databases for hidden impact vulnerabilities. We present a vulnerability analysis from January 2006 to April 2011 for two well known software packages: Linux kernel and MySQL. We show that 32% (Linux) and 62% (MySQL) of vulnerabilities discovered in this time period were hidden impact vulnerabilities. We also show that the percentage of hidden impact vulnerabilities has increased from 25% to 36% in Linux and from 59% to 65% in MySQL in the last two years. We then propose a hidden impact vulnerability identification methodology based on text mining classifier for bug databases. Finally, we discuss potential challenges faced by a development team when using such a classifier.

This paper presents a new dynamic programming method called the Iterative Learning Heuristic Dynamic Programming (ILHDP). The ILHDP is an Iterative Learning Control (ILC) based Neural Dynamic Programming (NDP) algorithm. The NDP aspect of the ILHDP algorithm is borrowed from traditional Adaptive Critic Design (ACD) algorithms. Typical NDP algorithms in the ACD class of algorithms train a Model Network beforehand and use a Critic Network, as the gradient approximator, trained back-and-forth with the Action Network in each iteration to converge the Action Network towards the optimal control policy. The proposed ILHDP algorithm updates the Model Network continually based on newly obtained data sampled during each Action Network optimization step on the same experiment. This process of Model Network updation ensures better gradient approximation presented by the Model Network itself. The presented ILHDP is used for the design of a Steam Power Plant controller with respect to the Active-Power-to-Frequency droop characteristics. Test results indicated that the ILHDP designed controller was capable of stabilizing the output power of the Steam Power Plant to track the load with a maximum tracking error of 0.011 for abrupt load changes as fast as 15s. The Steam Power Plant was also subjected to large transient spikes for which the designed controller proved to recover the system back to stability.

This paper presents the development of a new neural network based optimizer called the Adaptive Critic Learning Agent (ACLA) algorithm. The ACLA algorithm is based on the traditional Adaptive Critic Design (ACD) algorithm and hence its name. Conventional neural network based optimizers use the principle of Hopfield/Tank Neural Networks (HTNN) to solve unimodal optimization problems. These neural networks require tailored structures for the specific optimization problem. The ACLA algorithm presented in this paper uses a general randomly initialized neural network to solve any unimodal optimization problem. This is achieved by extending the principles of the traditional ACD algorithm for the ACLA algorithm. Other attributes of the ACLA algorithm are related to the issues with swarm based optimizers such as Particle Swarm Optimization (PSO) and Genetic Algorithms (GA). These issues are: (1) large memory requirements and (2) multiple parameters required to tune the algorithm's convergence performance. The ACLA algorithm resolves these issues by: (1) using only one neuron to reduce memory requirements and (2) using only a single learning coefficient parameter to tune the algorithm's convergence performance. The ACLA algorithm was tested and compared with three swarm based optimizers on two unimodal benchmark problems typically used for PSO and GA algorithms. Test results proved the ACLA algorithm to converge to solutions 7 orders greater than the swarm based algorithms. The ACLA algorithm was further tested on two multimodal benchmark problems to demonstrate its capability to converge to nearest local minima.

Data mining methods are becoming vital as the amount and complexity of available data is rapidly growing. Visual data mining methods aim at including a human observer in the loop and leveraging human perception for knowledge extraction. However, for large datasets, the rough knowledge gained via visualization is often times not sufficient. Thus, in such cases data summarization can provide a further insight into the problem at hand. Linguistic descriptors such as linguistic summaries and linguistic rules can be used in data summarization to further increase the understandability of datasets. This paper presents a Visual Linguistic Summarization tool (VLS-SOM) that combines the visual data mining capability of the Self-Organizing Map (SOM) with the understandability of linguistic descriptors. This paper also presents new quality measures for ranking of predictive rules. The presented data mining tool enables users to 1) interactively derive summaries and rules about interesting behaviors of the data visualized though the SOM, 2) visualize linguistic descriptors and visually assess the importance of generated summaries and rules. The data mining tool was tested on two benchmark problems. The tool was helpful in identifying important features of the datasets. The visualization enabled the identification of the most important summaries. For classification, the visualization proved useful in identifying multiple rules that classify the dataset.

Recently, type-2 fuzzy logic systems (T2 FLSs) have received increased research attention due to their potential to model and cope with the dynamic uncertainties ubiquitous in many engineering applications. However, because of the complex nature and the computational intensity of the inference process, only the constrained version of T2 FLSs, i.e., the interval T2 FLSs, was typically used. Fortunately, the very recently introduced concepts of α-planes and zSlices allow for efficient representation, as well as a computationally fast inference process, with general T2 (GT2) FLSs. This paper addresses the type-reduction phase in GT2 FLSs, using GT2 fuzzy sets (FSs) represented in the α-plane framework. The monotone property of centroids of a set of α-planes is derived and leveraged toward developing a simple to implement but fast algorithm for type reduction of GT2 FSs - i.e., the monotone centroid flow (MCF) algorithm. When compared with the centroid flow (CF) algorithm, which was previously developed by Zhai and Mendel, the MCF algorithm features the following advantages. 1) The MCF algorithm computes numerically identical centroid as the Karnik-Mendel (KM) iterative algorithms, unlike the approximated centroid which is obtained with the CF algorithm; 2) the MCF algorithm is faster than the CF algorithm, as well as the independent application of the KM algorithms; 3) the MCF algorithm is easy to implement, unlike the CF algorithm, which requires computation of the derivatives of the centroid; and 4) the MCF algorithm completely eliminates the need to apply the KM iterative procedure to any α-planes of the GT2 FS. The performance of the algorithm is presented on benchmark problems and compared with other type-reduction techniques that are available in the literature.

Pattern recognition in real-world data is subject to various sources of uncertainty that should be appropriately managed. The focus of this paper is the management of uncertainty associated with parameters of fuzzy clustering algorithms. Type-2 fuzzy sets (T2 FSs) have received increased research interest over the past decade, primarily due to their potential to model various uncertainties. However, because of the computational intensity of the processing of general T2 fuzzy sets (GT2 FSs), only their constrained version, i.e., the interval T2 (IT2) FSs, were typically used. Fortunately, the recently introduced concepts of α-planes and zSlices allow for efficient representation and computation with GT2 FSs. Following this recent development, this paper presents a novel approach for uncertain fuzzy clustering using the general type-2 fuzzy C-means (GT2 FCM) algorithm. The proposed method builds on top of the previously published IT2 FCM algorithm, which is extended via the α- planes representation theorem. The fuzzifier parameter of the FCM algorithm can be expressed using linguistic terms such as “small” or “high,” which are modeled as T1 FSs. This linguistic fuzzifier value is then used to construct the GT2 FCM cluster membership functions. The linguistic uncertainty is transformed into uncertain fuzzy positions of the extracted clusters. The GT2 FCM algorithm was found to balance the performance of T1 FCM algorithms in various uncertain pattern recognition tasks and to provide increased robustness in situations where noisy or insufficient training data are present.

Resiliency and security in critical infrastructure control systems in the modern world of cyber terrorism constitute a relevant concern. Developing a network security system specifically tailored to the requirements of such critical assets is of a primary importance. This paper proposes a novel learning algorithm for anomaly based network security cyber sensor together with its hardware implementation. The presented learning algorithm constructs a fuzzy logic rule base modeling the normal network behavior. Individual fuzzy rules are extracted directly from the stream of incoming packets using an online clustering algorithm. This learning algorithm was specifically developed to comply with the constrained computational requirements of low-cost embedded network security cyber sensors. The performance of the system was evaluated on a set of network data recorded from an experimental test-bed mimicking the environment of a critical infrastructure control system.

Many computational intelligence techniques for anomaly based network intrusion detection can be found in literature. Translating a newly discovered intrusion recognition criteria into a distributable rule can be a human intensive effort. This paper explores a multi-modal genetic algorithm solution for autonomous rule creation. This algorithm focuses on the process of creating rules once an intrusion has been identified, rather than the evolution of rules to provide a solution for intrusion detection. The algorithm was demonstrated on anomalous ICMP network packets (input) and Snort rules (output of the algorithm). Output rules were sorted according to a fitness value and any duplicates were removed. The experimental results on ten test cases demonstrated a 100 percent rule alert rate. Out of 33,804 test packets 3 produced false positives. Each test case produced a minimum of three rule variations that could be used as candidates for a production system.

Recently, Type-2 (T2) Fuzzy Logic Systems (FLSs) gained increased attention due to their capability to better describe, model and cope with the ubiquitous dynamic uncertainties in many engineering applications. By far the most widely used type of T2 FLSs are the Interval T2 (IT2) FLSs. This paper provides a comparative analysis of two fundamentally different approaches to defuzzification of IT2 Fuzzy Sets (FSs) the deterministic Karnik-Mendel Iterative Procedure (KMIP) and the stochastic sampling defuzzifier. As previously demonstrated by other researchers, these defuzzification algorithms do not always compute identical output values. In the presented work, the concept of centroid density of an IT2 FS is introduced in order to explain such discrepancies. It was demonstrated that the stochastic sampling defuzzification method converges towards the center of gravity of the proposed centroid density function. On the other hand, the KMIP method calculates the midpoint of the interval centroid obtained according to the extension principle. Since the information about the centroid density is removed via application of the extension principle, the two methods produce inevitably different results. As further demonstrated, this difference significantly increases in case of non-symmetric IT2 FSs.

Current developments in autonomous vehicle systems typically consider solutions to single problems like road detection, road following and object recognition individually. The integration of these individual systems into a single package becomes difficult because they are less compatible. This paper introduces a generic Integrated Neural System for Autonomous Vehicles (INSAV) package solution with processing blocks that are compatible with each other and are also suitable for hardware implementation. The generic INSAV is designed to account for important problems such as road detection, road structure learning, path tracking and obstacle detection. The paper begins the design of the generic INSAV by building its two most important blocks: the Road Structuring and Path Tracking Blocks. The obtained results from implementing the two blocks demonstrate an average of 92% accuracy of segmenting the road from a given image frame and path tracking of straight roads for stable motion and obstacle detection.

Data mining techniques are becoming indispensable as the amount and complexity of available data is rapidly growing. Visual data mining techniques attempt to include a human observer in the loop and leverage human perception for knowledge extraction. This is commonly allowed by performing a dimensionality reduction into a visually easy-to-perceive 2D space, which might result in significant loss of important spatial and topological information. To address this issue, this paper presents the design and implementation of a unique 3D visual data mining framework - CAVE-SOM. The CAVE-SOM system couples the Self-Organizing Map (SOM) algorithm with the immersive Cave Automated Virtual Environment (CAVE). The main advantages of the CAVE-SOM system are: i) utilizing a 3D SOM to perform dimensionality reduction of large multi-dimensional datasets, ii) immersive visualization of the trained 3D SOM, iii) ability to explore and interact with the multi-dimensional data in an intuitive and natural way. The CAVE-SOM system uses multiple visualization modes to guide the visual data mining process, for instance the data histograms, U-matrix, connections, separations, uniqueness and the input space view. The implemented CAVE-SOM framework was validated on several benchmark problems and then successfully applied to analysis of wind-power generation data. The knowledge extracted using the CAVE-SOM system can be used for further informed decision making and machine learning.

Type-2 Fuzzy Logic Controllers (T2 FLCs) have been recently applied in many engineering areas. While understanding the control potentials of T2 FLCs can still be considered an open question researchers, commonly claim superiority of T2 FLCs based on a limited exploration of the space of design parameters. The contribution of this work is based on a problem-driven design of uncertainty-robust Interval T2 (IT2) FLCs. The presented methodology starts with a baseline optimized T1 FLC. Next, a group of IT2 FLCs is designed using partially dependent approach by symmetrically blurring the membership functions around the original T1 fuzzy sets. This constrained design space allows for its systematic exploration and analysis. The performance of the designed controllers was evaluated on delta parallel robot hardware under two kinds of commonly encountered uncertainties: i) sensory noise and ii) uncertain system parameters. The experimental results showed that IT2 FLCs provide improved control performance against T1 FLCs when appropriate design of IT2 fuzzy sets is performed. In addition, it was demonstrated that excessive amount of “type-2 fuzziness” in the IT2 FLC design leads to rapid performance degradation.

Due to modern pervasive wireless technologies and high-performance monitoring systems, spatio-temporal information plays an important role in areas such as intelligent transportation systems (ITS), surveillance, scheduling, planning, or industrial automation. Security or criminal/terrorist threat prevention in modern ITS is one of today's most relevant concerns. This paper presents an algorithm for online spatio-temporal risk assessment in urban environments. In its first phase, the algorithm uses the online nearest neighbor clustering (NNC) algorithm to identify a set of significant places. In the second phase, a fuzzy inference engine is employed to quantify the level of risk that each significant place poses to the place of interest (e.g., vehicle, person, building, or an object of high assets). The contributions of the presented algorithm are given as follows: 1) recognition and extraction of the set of the most significant places; 2) dynamic adaptation of the solution to time-dependent traffic distributions; 3) parametric control by adjusting geographical proximity threshold, significance threshold, and discount factor; and 4) online risk assessment. The performance of the algorithm was demonstrated on a problem of traffic density estimation and risk assessment in a virtual urban environment.

General type-2 fuzzy logic systems (T2 FLS) constitute a powerful tool for coping with ubiquitous uncertainty in many engineering applications. However, the immense computational complexity associated with defuzzification of general T2 fuzzy sets still remains an unresolved issue and prohibits its practical use. This paper proposes a novel importance sampling based defuzzification method for general T2 FLS. Here, a subset from the domain of all embedded fuzzy sets is randomly sampled using a specific probability distribution function. The algorithm is compared with the previously published uniform sampling defuzzification method. Experimental results demonstrate that importance sampling substantially reduces the variance of the sampling defuzzification method. Comparison of T2FLS output surfaces showed that smoother and more stable response can be achieved with the proposed importance sampling based defuzzification method.

When training a neural network it is tempting to experiment with architectures until a low total error is achieved. The danger in doing so is the creation of a network that loses generality by over-learning the training data; lower total error does not necessarily translate into a low total error in validation. The resulting network may keenly detect the samples used to train it, without being able to detect subtle variations in new data. In this paper, a method is presented for choosing the best neural network architecture for a given data set based on observation of its accuracy, precision, and mean square error. The method, based on [1], relies on k-fold cross validation to evaluate each network architecture k times to improve the reliability of the choice of the optimal architecture. The need for four separate divisions of the data set is demonstrated (testing, training, and validation, as normal, and an comparison set). Instead of measuring simply the total error the resulting discrete measures of accuracy, precision, false positive, and false negative are used. This method is then applied to the problem of locating cryptographic algorithms in compiled object code for two different CPU architectures to demonstrate the suitability of the method.

Detecting an aircraft solely by its infrared (IR) signature in real time can be extremely challenging task depending on the image background clutter. Neural networks offer a reliable method of detecting targets (aircraft) against a multitude of background scenes and a variety of environmental conditions. Neural networks can rapidly "learn" to differentiate between background clutter and fast moving, small, "hot" (temperature) targets. A neural network real time video processor (NN-RTVP) presented in this paper was inspired by and a Kohonen neural network (KNN) approach to not only process "still" frames but also process video in real time. Experimental results demonstrated that it is possible to provide real time "point-outs" of thermally significant objects.

Finding and identifying cryptography is a growing concern in the malware analysis community. In this paper, artificial neural networks are used to classify functional blocks from a disassembled program as being either cryptography related or not. The resulting system, referred to as NNLC (neural net for locating cryptography) is presented and results of applying this system to various libraries are described.

Support vector machines (SVMs) represent a well known technique for data classification. However, the complexity of the training process makes the SVMs unsuitable for classifying large datasets. Examples of existing approaches to this problem are sampling of the input datasets or clustering of similar inputs. On the other hand, the growing neural gas algorithm (GNG) is a robust tool for cluster analysis, capable of learning the topology of the data. It overcomes most of the common issues of clustering techniques such as predefined number of clusters or beforehand specified cluster radius. This paper presents a solution to the problem of classifying large datasets via learning of the data topology. The described algorithm combines the GNG algorithm with the SVM solver into a specific algorithm for classification of large datasets - the GNG-SVM framework. The input dataset is first preprocessed with the GNG algorithm. A new reduced training dataset is created from the extracted topological knowledge. Because the size of the dataset is significantly reduced, the training process of the SVM solver becomes substantially less memory demanding. The performance of the proposed GNG-SVM framework is tested on both synthetic and benchmark real world datasets.

Resiliency and security in control systems such as SCADA and nuclear plant's in today's world of hackers and malware are a relevant concern. Computer systems used within critical infrastructures to control physical functions are not immune to the threat of cyber attacks and may be potentially vulnerable. Tailoring an intrusion detection system to the specifics of critical infrastructures can significantly improve the security of such systems. The IDS-NNM - intrusion detection system using neural network based modeling, is presented in this paper. The main contributions of this work are: 1) the use and analyses of real network data (data recorded from an existing critical infrastructure); 2) the development of a specific window based feature extraction technique; 3) the construction of training dataset using randomly generated intrusion vectors; 4) the use of a combination of two neural network learning algorithms - the error-back propagation and Levenberg-Marquardt, for normal behavior modeling. The presented algorithm was evaluated on previously unseen network data. The IDS-NNM algorithm proved to be capable of capturing all intrusion attempts presented in the network communication while not generating any false alerts.

Time and location data play a very significant role in a variety of factory automation scenarios, such as automated vehicles and robots, their navigation, tracking, and monitoring, to services of optimization and security. Pervasive wireless capabilities combined with time and location information are enabling new applications in areas such as transportation systems, health care, elder care, military, emergency response, critical infrastructure, and law enforcement. A wireless object in proximity to some area for a duration of time may pose a risk hazard to the environment. This paper presents a novel fuzzy based spatio-temporal risk calculation DSTiPE method that a wireless object may present to the environment. The presented Matlab based application for cluster extraction is verified on a diagonal vehicle movement example.

In factory automation, production line scheduling entails a number of competing issues. Finding optimal configurations often requires use of local search techniques. Local search looks for a goal state employing heuristics and random local ldquoprobesrdquo in order to move from state to state. All local search techniques, however, suffer from problems with local maxima, i.e. have the potential of getting ldquostuckrdquo in a suboptimal state. While careful introduction of randomizations is certainly a recognized technique, it can also lead the algorithm even more astray. This paper describes a heuristic technique called descending deviation optimizations (DDO) in which a gradually lowering-randomization ceiling allows a local search technique to ldquobouncerdquo randomly without going too far astray. An example applying the DDO to a local search technique and achieving significant improvement is shown.

Wireless technology is a central component of many factory automation infrastructures in both the commercial and government sectors, providing connectivity among various components in industrial realms (distributed sensors, machines, mobile process controllers). However wireless technologies provide more threats to computer security than wired environments. The advantageous features of Bluetooth technology resulted in Bluetooth units shipments climbing to five million per week at the end of 2005 [1, 2]. This is why the real-time interpretation and understanding of Bluetooth traffic behavior is critical in both maintaining the integrity of computer systems and increasing the efficient use of this technology in control type applications. Although neuro-fuzzy approaches have been applied to wireless 802.11 behavior analysis in the past, a significantly different Bluetooth protocol framework has not been extensively explored using this technology. This paper presents a new neuro-fuzzy traffic analysis algorithm of this still new territory of Bluetooth traffic. Further enhancements of this algorithm are presented along with the comparison against the traditional, numerical approach. Through test examples, interesting Bluetooth traffic behavior characteristics were captured, and the comparative elegance of this computationally inexpensive approach was demonstrated. This analysis can be used to provide directions for future development and use of this prevailing technology in various control type applications, as well as making the use of it more secure.

The redundancy regeneration (sparing or rebuild) algorithms in disk arrays face the problem of balancing between the data recovery activity within the array and the user workload acting upon the array at the same time [1]. If the algorithm favors the user workload so the user requests can always preempt the internal data recovery, then the data sparing can stall in the presence of a sustained workload. But on the contrary, if the data recovery is favored over the user requests, the latency of the user requests can be so high to reach unacceptable levels for the data transactions. Using computationally intelligent techniques, like fuzzy logic, better algorithms to balance the level of user requests and the internal data recovery can be achieved. The disk array and data recovery process are modeled using the queue systems with vacations (QSV) [2]. A fuzzy algorithm to control the sparing is presented in this paper. The results indicate that by using fuzzy logic, a better balancing is achieved between the need to have an acceptable response time for the user requests and the data recovered as soon as possible.

The performability of disk arrays systems has been studied before. However, in the case of imprecise data, a fuzzy model can be the base for the performability analysis. This paper presents a performability analysis of an MSExchange-like e-mail server. The analysis is based on a Markov reward model. The performability analysis is accomplished through the use of fuzzy arithmetic. Unlike traditional Markov chains, fuzzy Markov chains can successfully handle uncertain, imprecise probabilities. In cases where the failure rates, repair rates, or the workload parameters are uncertain, Markov Chains enhanced with fuzzy arithmetic provide means for comprehensive predictive performability analysis of a system. This performability analysis provides a valuable guideline regarding required resources such as the number of mailboxes, and therefore, the number of users the mail server can support with regards to the reliability and performance of the disk array used by the mail server. The fuzzy arithmetic helps in better visualization and estimation of the range of number of users the mail server is capable of servicing over long periods of time.