This study investigates the vulnerabilities of autonomous navigation and landing systems in Urban Air Mobility (UAM) vehicles. Specifically, it focuses on Trojan attacks that target deep learning models, such as Convolutional Neural Networks (CNNs). Trojan attacks work by embedding covert triggers within a model’s training data. These triggers cause specific failures under certain conditions, while the model continues to perform normally in other situations.We assessed the vulnerability of Urban Autonomous Aerial Vehicles (UAAVs) using the DroNet framework. Our experiments showed a significant drop in accuracy, from 96.4% on clean data to 73.3% on data triggered by Trojan attacks. To conduct this study, we collected a custom dataset and trained models to simulate real-world conditions. We also developed an evaluation framework designed to identify Trojan-infected models. This work demonstrates the potential security risks posed by Trojan attacks and lays the groundwork for future research on enhancing the resilience of UAM systems.

In this paper, we propose a conflict-free multi-agent flight scheduling that ensures robust separation in constrained airspace for Urban Air Mobility (UAM) operations application. First, we introduce Pairwise Conflict Avoidance (PCA) based on delayed departures, leveraging kinematic principles to maintain safe distances. Next, we expand PCA to multi-agent scenarios, formulating an optimization approach that systematically determines departure times under increasing traffic densities. Performance metrics, such as average delay, assess the effectiveness of our solution. Through numerical simulations across diverse multi-agent environments and real-world UAM use cases, our method demonstrates a significant reduction in total delay while ensuring collision-free operations. This approach provides a scalable framework for emerging urban air mobility systems.

As autonomous vehicles become an essential component of modern transportation, they are increasingly vulnerable to threats such as GPS spoofing attacks. This study presents an adaptive detection approach utilizing a dynamically tuned Density Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, designed to adjust the detection threshold (ε) in real-time. The threshold is updated based on the recursive mean and standard deviation of displacement errors between GPS and in-vehicle sensors data, but only at instances classified as non-anomalous. Furthermore, an initial threshold, determined from 120,000 clean data samples, ensures the capability to identify even subtle and gradual GPS spoofing attempts from the beginning. To assess the performance of the proposed method, five different subsets from the real-world Honda Research Institute Driving Dataset (HDD) are selected to simulate both large and small magnitude GPS spoofing attacks. The modified algorithm effectively identifies turn-by-turn, stop, overshoot, and multiple small biased spoofing attacks, achieving detection accuracies of 98.62±1%, 99.96±0.1%, 99.88±0.1%, and 98.38±0.1%, respectively. This work provides a substantial advancement in enhancing the security and safety of AVs against GPS spoofing threats.

The integration of Unmanned Aircraft Systems (UAS) into the National Airspace System (NAS) requires robust encounter modeling tools to evaluate Detect-and-Avoid (DAA) systems. However, existing tools often lack the ability to model the unique dynamics of large UAS (lUAS) and small UAS (sUAS), and few are available as open-source solutions. In this paper, we introduce novel flight modeling concepts for lUAS and sUAS, developed within an open-source framework. For lUAS, we propose a hybrid modeling strategy that combines probabilistic manned aircraft models with UAS-specific performance constraints. A machine learning-based surrogate model is employed to streamline feasibility evaluation and enable the generation of realistic trajectories. The sUAS flight modeling technique enables mission-aware trajectory generation by incorporating geospatial data and customized control architecture for fixed-wing and multirotor configurations. These models capture a wide range of aircraft dynamics, offering the potential to generate versatile encounter datasets for DAA evaluation. By releasing them as open-source resources, we aim to encourage broader collaboration, inform regulatory development, and drive innovation in encounter modeling, all in support of the safe and effective integration of UAS into the NAS.

Unmanned Aerial Vehicles (UAVs) are increasingly employed in cooperative surveillance missions where data collection across disparate areas is crucial. In such systems, data from all UAVs is collected and processed in a central server, making it vulnerable to breaches and unauthorized access. Federated Learning (FL) addresses these concerns by enabling collaborative model training without centralized data collection. In FL, each UAV trains a local model on its own data and only shares the model updates with a central server. The central server then aggregates these parameters to update a global model, which is redistributed to all participating UAVs. However, FL’s reliance on a central server introduces challenges, especially in geographically highly dynamic and dispersed scenarios. The central server can become a single point of failure and may struggle with the communication overhead and latency issues inherent in such dynamic environments. To overcome these limitations, this paper proposes a decentralized federated learning framework for multi-agent UAV systems. This approach facilitates direct sharing of local deep learning (DL) model parameters among UAVs, eliminating the need for a central server. Our approach employs the Metropolis-Hastings algorithm to ensure UAVs achieve consensus on shared model parameters, ensuring balanced weight distribution and stable training processes. We validate our fully distributed DL model aggregation using the ResNet-18 model. Our results confirm DFL’s effectiveness in achieving low RMSE values and rapid convergence, comparable to centralized FL, across dynamic UAV networks.

This paper proposes a novel data-driven technique, namely Online Learning for sensor Fault diagnosis Analysis (OLFA), to perform real-time fault analysis for autonomous cars. Considering the non-stationary properties of real-time sensor faults and the mapping relationship between sensors and feature variables, the proposed method decomposes the sensor fault diagnosis analysis problem into an online data stream classification and feature ranking problems. To detect and identify faults, a clustering-based data stream classification approach is developed to continuously capture and classify non-stationary sensor faults for autonomous cars with little intervention from human experts. An effective active learning method is extended and embedded into the proposed framework to minimize the need for prior knowledge about faults and enable the continual learning capability to adapt to and handle the non-stationary properties of sensor faults. Moreover, the proposed framework addresses the parameter optimization issue of existing machine learning based fault analysis techniques and employs feature ranking analysis to systematically analyze the possible source(s) of sensor faults. CAR Learning to Act (CARLA), a well-known realistic autonomous driving simulator, is used as the benchmark to perform the sensor fault injection and online data stream collection to evaluate the efficacy of OLFA. Analysis of the collected faulty datasets and experimental results, and comparison between OLFA and several state-of-the-art clustering-based approaches for fault classification, demonstrated the efficacy of the proposed framework in the domain of autonomous cars.

The growth in wireless traffic and mobility of devices have congested the core network significantly. This bottleneck, along with spectrum scarcity, made the conventional cellular networks insufficient for the dissemination of large contents. The ability of content-centric networking (CCN) and device-to-device (D2D) communication in offloading the network and huge unlicensed spectrum at millimeter-wave (mmWave) band, make the integration of CCN with D2D communication in the mmWave band a viable solution to improve the network's throughput. In this paper, we propose a novel scheme that enables efficient initialization of CCN-based D2D networks in the mmWave band through addressing decentralized D2D peer association and antenna beamwidth selection. The proposed scheme considers mmWave characteristics such as directional communication and blockage susceptibility. We propose a heuristic peer association algorithm to associate D2D users using context information, including link stability time and content availability. We model the beamwidth selection problem as a potential game and propose a synchronous log-linear learning algorithm to obtain the game's optimal Nash equilibrium. The performance of the proposed scheme in terms of data throughput and transmission efficiency is evaluated through extensive simulations. Simulation results show that the proposed scheme improves network performance significantly and outperforms other methods in the literature.

Increased traffic density with a greater degree of increased automation in aviation is expected within the next decade. Therefore, airspace capacity will become more congested and result in increasing challenges for detecting conflicts between aerial vehicles. Furthermore, because these vehicles rely on surrounding vehicles following a planned path, it is essential to identify flights not following a planned direction. In this paper, we utilize an ensemble of the existing outlier detection approaches for identifying the anomalous flight trajectories. In the initial step, flight trajectories are preprocessed to extract and process vital features, with the next step of having twenty different outlier detection algorithms assembled to classify trajectories. Throughout our extensive experiments and comparison studies, promising results are shown including the effectiveness of different anomaly detection algorithms and how utilizing feature engineering can improve the results of these outlier detection methods.

Realtime estimation of travel time is a key traffic parameter for designing and planning for transportation systems, particularly when providing mobility-on-demand (MOD) services. However, the analysis and prediction of travel time can be delayed significantly due to the complexity and huge computational requirements of microsimulation models. Thus, as an alternative solution, we propose a data-driven approach for the efficient and reliable prediction of travel time. Our approach takes advantage of the strengths of SVM and ARIMA for fully capturing the traffic patterns in the traffic data. We introduce a new parameter $\kappa$ into the SVM-ARIMA model to adjust the weight of the ARIMA component, which significantly improves the performance. We validate the performance of the proposed approach using data generated from a microsimulation platform. Our experimental results and comparisons with the existing ML-based methods demonstrates the efficacy of the proposed data-driven approach.

The purpose of this paper is to devise an interpretable hybrid classification model for Convolutional Neural Networks (CNN) and a Learning Classifier System (LCS). The presented hybrid system integrates the fundamental attributes from both types of these classifiers. In the proposed hybrid model CNN works as an automatic feature extractor, and LCS works to provide interpretable rule-based classification results. Although LCS has limitations working on higher dimensional datasets, we resolve this limitation by using CNN as a feature extractor. The other concept of the non-interpretability of CNN is addressed by using the LCS rule. Furthermore, our experiment with higher dimensional datasets like CIFAR-10 and Fashion-MNIST shows that extended LCS provides comparable performance to the standard neural network model while also providing interpretable results. We named this extended LCS method Convolutional Learning Classifier Cystem (C-LCS).

As complex autonomous systems become increasingly ubiquitous, their deployment and integration into our daily lives will become a significant endeavor. Human–machine trust relationship is now acknowledged as one of the primary aspects that characterize a successful integration. In the context of human–machine interaction (HMI), proper use of machines and autonomous systems depends both on the human and machine counterparts. On one hand, it depends on how well the human relies on the machine regarding the situation or task at hand based on willingness and experience. On the other hand, it depends on how well the machine carries out the task and how well it conveys important information on how the job is done. Furthermore, proper calibration of trust for effective HMI requires the factors affecting trust to be properly accounted for and their relative importance to be rightly quantified. In this article, the functional understanding of human–machine trust is viewed from two perspectives—human-centric and machine- centric. The human aspect of the discussion outlines factors, scales, and approaches, which are available to measure and calibrate human trust. The discussion on the machine aspect spans trustworthy artificial intelligence, built-in machine assurances, and ethical frameworks of trustworthy machines.

In Multi-label (ML) classification, each instance is associated with more than one class label. Incorporating the label correlations into the model is one of the increasingly studied areas in the last decade, mostly due to its potential in training more accurate predictive models and dealing with noisy/missing labels. Previously, multi-label learning classifier systems have been proposed that incorporate the high-order label correlations into the model through the label powerset (LP) technique. However, such a strategy cannot take advantage of the valuable statistical information in the label space to make more accurate inferences. Such information becomes even more crucial in ML image classification problems where the number of labels can be very large. In this paper, we propose a multi-label learning classifier system that leverages a structured representation for the labels through undirected graphs to utilize the label similarities when evolving rules. More specifically, we propose a novel scheme for covering classifier actions, as well as a method to calculate ML prediction arrays. The effectiveness of this method is demonstrated by experimenting on multiple benchmark datasets and comparing the results with multiple well-known ML classification algorithms.

The non-stationary nature of data streams strongly challenges traditional machine learning techniques. Although some solutions have been proposed to extend traditional machine learning techniques for handling data streams, these approaches either require an initial label set or rely on specialized design parameters. The overlap among classes and the labeling of data streams constitute other major challenges for classifying data streams. In this paper, we proposed a clustering-based data stream classification framework to handle non-stationary data streams without utilizing an initial label set. A density-based stream clustering procedure is used to capture novel concepts with a dynamic threshold and an effective active label querying strategy is introduced to continuously learn the new concepts from the data streams. The sub-cluster structure of each cluster is explored to handle the overlap among classes. Experimental results and quantitative comparison studies reveal that the proposed method provides statistically better or comparable performance than the existing methods.

Standard Convolutional Neural Network (CNN) designs rarely focus on the importance of explicitly capturing diverse features to enhance the network’s performance. Instead, most existing methods follow an indirect approach of increasing or tuning the networks’ depth and width, which in many cases significantly increase the computational cost. Inspired by biological visual system, we proposes a Diverse and Adaptive Attention Convolutional Network (DA2-Net), which enables any feed-forward CNNs to explicitly capture diverse features and adaptively select and emphasize the most informative features to efficiently boost the network’s performance. DA2-Net incurs negligible computational overhead and it is designed to be easily integrated with any CNN architecture. We extensively evaluated DA2-Net on benchmark datasets, including CIFAR100, SVHN, and ImageNet, with various CNN architectures. The experimental results show DA2-Net provides a significant performance improvement with very minimal computational overhead.

In this paper, we present a Robust Completed Local Binary Pattern (RCLBP) framework for a surface defect detection task. Our approach uses a combination of Non-Local (NL) means filter with wavelet thresholding and Completed Local Binary Pattern (CLBP) to extract robust features which are fed into classifiers for surface defects detection. This paper combines three components: A denoising technique based on Non-Local (NL) means filter with wavelet thresholding is established to denoise the noisy image while preserving the textures and edges. Second, discriminative features are extracted using the CLBP technique. Finally, the discriminative features are fed into the classifiers to build the detection model and evaluate the performance of the proposed framework. The performance of the defect detection models are evaluated using a real-world steel surface defect database from Northeastern University (NEU). Experimental results demonstrate that the proposed approach RCLBP is noise robust and can be applied for surface defect detection under varying conditions of intraclass and inter-class changes and with illumination changes.

The growth in the transportation sector has led to an enormous increase in the number of vehicles that ply our roads daily. Even though this advancement has provided numerous transportation modes, it has resulted in serious transportation issues including road congestion. Hence, estimating the number of vehicles on a road will enable traffic managers to take appropriate decisions to curb congestion. In this paper, we propose to use an extreme gradient boosting (XGBoost) algorithm to efficiently and accurately predict the hourly traffic volume. We investigate the effectiveness of the proposed method for different scenarios including how well it performs during extreme weather conditions and holidays. We further investigate the effect of ridge and LASSO regularization on the performance of XGBoost. We then propose a new approach for setting the LASSO regularization parameter in terms of the number of observations and predictors. The performance and computational efficiency of the proposed approach is evaluated on data collected from Interstate-94, Minnesota and the results are compared with existing methods. The results show that the proposed method provides a good balance between performance and computational efficiency.

Feature selection methods are widely used to address the high computational overheads and curse of dimensionality in classifying high-dimensional data. Most conventional feature selection methods focus on handling homogeneous features, while real-world datasets usually have a mixture of continuous and discrete features. Some recent mixed-type feature selection studies only select features with high relevance to class labels and ignore the redundancy among features. The determination of an appropriate feature subset is also a challenge. In this paper, a supervised feature selection method using density-based feature clustering (SFSDFC) is proposed to obtain an appropriate final feature subset for mixed-type data. SFSDFC decomposes the feature space into a set of disjoint feature clusters using a novel density-based clustering method. Then, an effective feature selection strategy is employed to obtain a subset of important features with minimal redundancy from those feature clusters. Extensive experiments as well as comparison studies with five state-of-the-art methods are conducted on SFSDFC using thirteen real-world benchmark datasets and results justify the efficacy of the SFSDFC method.

In this paper, a driver's intention prediction near a road intersection is proposed. Our approach uses a deep bidirectional Long Short-Term Memory (LSTM) with an attention mechanism model based on a hybrid-state system (HSS) framework. As intersection is considered to be as one of the major source of road accidents, predicting a driver's intention at an intersection is very crucial. Our method uses a sequence to sequence modeling with an attention mechanism to effectively exploit temporal information out of the time-series vehicular data including velocity and yaw-rate. The model then predicts ahead of time whether the target vehicle/driver will go straight, stop, or take right or left turn. The performance of the proposed approach is evaluated on a naturalistic driving dataset and results show that our method achieves high accuracy as well as outperforms other methods. The proposed solution is promising to be applied in advanced driver assistance systems (ADAS) and as part of active safety system of autonomous vehicles.

Despite advancements in vehicle security systems, over the last decade, auto-theft rates have increased, and cyber-security attacks on internet-connected and autonomous vehicles are becoming a new threat. In this paper, a deep learning model is proposed, which can identify drivers from their driving behaviors based on vehicle telematics data. The proposed Long-Short-Term-Memory (LSTM) model predicts the identity of the driver based on the individual's unique driving patterns learned from the vehicle telematics data. Given the telematics is time-series data, the problem is formulated as a time series prediction task to exploit the embedded sequential information. The performance of the proposed approach is evaluated on three naturalistic driving datasets, which gives high accuracy prediction results. The robustness of the model on noisy and anomalous data that is usually caused by sensor defects or environmental factors is also investigated. Results show that the proposed model prediction accuracy remains satisfactory and outperforms the other approaches despite the extent of anomalies and noise-induced in the data.

Connected and Automated Vehicles (CAV) are a great solution to many of the transportation problems such as congestion. Cooperative adaptive cruise control (CACC) is one of the CAV applications in which the vehicles are connected through wireless communication to coordinate their behavior in the form of a platoon. CACC has many advantages including an increase in safety and road capacity and a decrease in fuel consumption. However, wireless communications introduce network induced imperfections such as transmission delay and packet loss. In this study, the CACC system is modeled as a networked control system, and the impact of these communication uncertainties in individual links on the platoon formation and stability is investigated. The results suggest that packet loss even in one of the links can negatively affect string stability.

In this paper, a novel kinematic-based optimal trajectory planning formulation for an autonomous vehicle is presented. In this new formulation, the quadratic errors of position, velocity, and acceleration are minimized subject to the rear wheel car-like vehicle nonlinear kinematic model. Minimizing the error of velocity and acceleration in addition to the error of position, allows us to obtain both optimal vehicle trajectory and control law. The Variational approach is used to minimize the cost function. Then, optimal trajectory and control inputs are numerically calculated by solving a set of two-point boundary value (TPBV) nonlinear differential equations. Finally, the proposed method is evaluated in two scenarios of lane changing and multi-curvature road which verify the success of the proposed method in generating an optimal trajectory and control inputs.

Multi-sensor data fusion is widely used in various application domains. Integration of multiple sensors is a complex problem. This is because it is often characterized by uncertainty due to randomness and non-specificity. The Dempster Shafer (DS) theory of evidence has often been used for modelling and reasoning under uncertainty. However, the DS rule of combination is often prone to counter-intuitive results when combining pieces of evidence that are highly conflicting. As a result, several alternative combination rules have emerged. One approach is to assign weight to each basic probability assignment (BPA) prior to the use of the DS rule of combination. Most existing methods of assigning weight only focus on the credibility of each BPA without considering the reliability of the source of the BPA. In this work, we propose a multi-sensor data fusion that takes into consideration both the reliability of each BPA source and the credibility degree. A benchmark dataset was used to evaluate the effectiveness of the proposed method. To further assess the robustness of the proposed method in handling uncertainty, different noise levels were introduced to the training set.

When dealing with real-world problems, uncertainty in information is inevitable. Uncertainty can result from different sources such as noise, ambiguity or lack of knowledge. A category of problems withing the supervised learning framework in which the class membership of the training data is assessed by an expert and expressed in the form of a possibility distribution is considered in this paper. The problem is handled using the possibility theory framework. An accuracy-based learning classifier system designed for function approximation, namely XCSF is employed to model the possibility distribution of different classes. For unseen instances, the predictions from a set of matching rules are fused using an information fusion technique of possibility theory to create an aggregated distribution of possible classes. This distribution may then be either interpreted by a human operator to support decision-making or processed to obtain a final class prediction by selecting the class with the highest possibility. The experimental study with synthetic data reveals the ability of the proposed method to make efficient use of available information in the possibilistic labels of training data compared to conventional classification methods.

Multi-label classification deals with problem domains in which each instance belongs to more than one class simultaneously. Label Powerset (LP) is an efficient multi-label learning algorithm that considers each distinct combination of labels in training data as a unique new class and trains a conventional multi-class learning algorithm. In this paper a Multi-label classification algorithm is proposed that integrates LP with a rule-based evolutionary machine learning approach developed for supervised learning tasks, namely sUpervised Learning Classifiers (UCS). Moreover, to improve the prediction capability of the model on unseen instances, a prediction aggregation strategy is proposed to make efficient use of all the potentially helpful information in the rule base. The result is a multi-label rule-based evolutionary learner, which is called MLRBC (Multi-Label Rule-Based Classifier). Taking advantage of the strong generalization capability of UCS and its robustness in handling data sets with imbalanced classes, the proposed MLRBC algorithm is able to address some of the challenges involved in using LP. Experimental studies on multiple real-world datasets show that the proposed algorithm substantially improves the performance of the original LP technique and shows competitive performance against some of the state of the art multi-label learning algorithms.

As an effective dimensionality reduction technique, feature selection is widely used in the preprocessing procedure in data mining. It is highly advocated by its superiority in mitigating the effect of noisy data and simplifying the analysis of high-dimensional data. In this paper, a novel unsupervised feature selection procedure based on a clustering algorithm is proposed to evaluate the goodness of features and select a set of useful features without losing the characteristics of the data. It consists of two steps: clustering and feature evaluation. In the clustering procedure, a novel clustering algorithm based on the fitness proportionate sharing is adopted to separate data into distinct clusters without any prior knowledge about data, which is more applicable to the analysis of unknown datasets. On the other hand, the feature evaluation procedure will use the information extracted from the clustering procedure to evaluate the usefulness of each feature and select good features. The proposed method is simulated with four other famous existing feature selection algorithms and a comparison is provided in this paper. Simulation results on both synthetic and real datasets demonstrate that the proposed procedure of feature selection can effectively evaluate the significance of features and obtain a better subset of features than other four existing algorithms.

In this work the Michigan style strength-based learning classifier system, which is a rule-based supervised learning algorithm, is extended to handle multi-label classification tasks. Moreover, it is assumed that the class membership for training data is partially known and the uncertainty is represented by confidence values that reflects the probability of each label being true. Necessary parameters are introduced, and learning classifiers are modified to learn simultaneously the confidence level and multi-label in the training data. Additionally, to quantify the classifier performance, a novel loss measure is introduced that generalizes the well-known Hamming loss criteria to takes into account the classification error and confidence estimation error simultaneously. The algorithm is tested on one real-world data and two synthetic data sets. Results show the ability of the model in learning multi-class and multi-label data with low confidence estimation error.

A control approach for automated highway driving is proposed in this study, which can learn from human driving data, and is applied to the longitudinal trajectory of an autonomous car. Naturalistic driving data are used as samples to train the model offline. Then, the model is used online to emulate what a human driver would do by computing acceleration. This reference acceleration is tracked by a predictive controller, which enforces a set of comfort and safety constraints before applying the final acceleration. The controller is designed to balance between maintaining vehicle safety and following the model's commands. Thus, the proposed controller can handle dynamic traffic situations while performing like a human driver. This approach is validated on two different scenarios using MATLAB simulations.

Collision avoidance is an essential component in advanced driving assistance systems, as it ensures the safety of the vehicle in near crash or crash scenarios. In this study, a collision avoidance system for lane change events is proposed which plans the trajectory based on the level of danger. The danger level is computed by a fuzzy inference system developed with naturalistic driving data to better capture the real-world factors, which may cause an accident. In addition, a fault determination classifier is introduced in order to determine the responsible driver in a near crash event. This system is evaluated on simulated naturalistic near crash events and the results demonstrate good performance of the proposed system.

In this paper, driver intention estimation near a road intersection is presented, using discrete hidden Markov models (HMM) and the Hybrid State System (HSS) framework as basis. The development of Advanced Driver Assistance Systems (ADAS) has assisted drivers in many driving scenarios and resulted in safe driving. Developing techniques to estimate driver's intention leads to the advancement of ADAS. As a large number of accidents occur near road intersections, estimating the intention of a driver at an intersection is vital. The methods developed in this paper can be applied in ADAS to take appropriate measures in reducing accidents. The driver decisions are depicted as a Discrete State System (DSS) at a higher level and the continuous vehicle dynamics are depicted as a Continuous State System (CSS) at a lower level in the HSS framework. In the proposed technique, the vehicle's continuous observations including speed and yaw-rate, are used to estimate the driver's intention at each time step. In this work, the speed and yaw-rate are discretized in such a way that the important features about the driver's intention such as "go straight," "stop," "turn right," or "turn left" at the intersection, are abstractedly represented in the form of symbols. Naturalistic driving data, which is collected using a vehicle fitted with sensors, is used to train and test the developed model. The results from the proposed approach show high accuracy in estimating the driver's intention at a road intersection.

Existing clustering techniques primarily rely on prior knowledge about the data, such as the number of clusters and radii. However, in real applications, the number of clusters and the radii of clusters are usually unknown. Therefore, the performance of clustering methods with overlapping data is degraded due to their limitations in finding all cluster centers with uneven density values. Hence, a new clustering algorithm based on fitness proportionate sharing is proposed to map the problem into a multimodal optimization problem. In this paper, clusters are considered as niches, and the individuals with the highest density values of each niche are the cluster centers. Instead of using the traditional sharing strategy, the fitness proportionate sharing strategy is implemented in the identification of niche maxima to overcome the sensitivity of uneven density values of cluster centers. A procedure of niche expansion is employed for the merging of clusters. Simulation results and complexity analysis reveal that the proposed clustering algorithm based on fitness proportionate sharing provides a higher accuracy performance without any prior information.

Hurricanes constitute major natural disasters that lead to destruction and loss of lives. Therefore, to reduce economic loss and to save human lives, an accurate forecast of hurricane occurrences is crucial. Despite the availability of data and advanced forecasting techniques, there is a need for effective methods with higher accuracy of prediction. We propose a sparse Recurrent Neural Network (RNN) with flexible topology for trajectory prediction of the Atlantic hurricanes. Topology of the RNN along with the strength of the connections are evolved by a customized Genetic Algorithm. The network is particularly suitable for modeling of hurricanes which have complex systems with unknown dynamics. For prediction of the future trajectories of a target hurricane, the Dynamic Time Warping (DTW) distances between direction of the target hurricane over time, and other hurricanes in the dataset are determined and compared. The most similar hurricanes to the target hurricane are then used for training of the network. Comparisons between the actual tracks of the hurricanes DEAN, SANDY, ISSAC and HUMBERTO, and the generated predictions by the sparse RNN for one and two steps ahead of time show that our approach is quite promising for this aim.

Advanced driver assistance systems are required to detect latent hazards posed by surrounding vehicles and generate an appropriate response to enhance safety. Lane changes constitute potentially risky maneuvers, as drivers involved encounter latent hazards due to surrounding vehicles. A careful study of lane change behavior is therefore essential in identifying potential abnormalities that may lead to various hazards, during the process of a lane change. In this study, an anomaly detection technique is used to compare snapshots of normal and dangerous lane change maneuvers, to identify the abnormal instances. A one-class support vector machine is used and tested for novelty identification of naturalistic driving study data. The results show that the technique is able to detect dangerous lane changes with high accuracy. In addition, results suggest that dangerous behavior could occur before, after or during a lane change maneuver.

This study considers offline identification of switched linear MIMO systems using measurements from their inputs and outputs. This is a class of non-convex optimization and ill-posed problems. To convert this optimization into a binary integer programming problem, the proposed approach assumes that the number of switches among the subsystems is upper-bounded. The state-space realization of each sub-system is found by the subspace identification. The proposed approach does not need the tuning of the moving window size or any penalization factor. The algorithm efficiency is evaluated through numerical simulations. The results indicate that the error of identification is small and the eigenvalues of sub-systems are estimated successfully.

This paper addresses the motion planning problem for two autonomous robots (the defender and the attacker) with competitive objectives, which are involved in a reach-avoid scenario. This adversarial aspect of the game makes the problem complex with high computational cost. To address this problem, we propose a novel symbolic approach for the robot motion planning and control of the robots, which can effectively manage the complexity of the problem. The basic idea is to partition the environment into convex regions, and then, capture the desired objectives of the defender and the adversarial behavior of the attacker with temporal logic formulas. We also use finite two-player zero-sum games as a tool for the robot decision-making over the partitioned space. An illustrative examples has been provided to detail the steps of the proposed algorithm and the simulation results are presented to verify the effectiveness of the proposed algorithm.

The Smart home has gained widespread attention due to its flexible integration into everyday life. This next generation green home system, transparently unifies various home appliances, smart sensors and wireless communication technologies. It can integrate diversified physical sensed information and control various consumer home devices, with the support of active sensor networks having both sensor and actuator components. Although smart homes are gaining popularity due to their energy saving and better living benefits, there is no standardized design for smart homes. In this paper, we put forward a concept by designing and implementing a smart home system which can classify and predict the state of the home based on historical data. We set up a wireless sensor network and collected months of data. By employing supervised machine learning technique, we were able to establish patterns and use the acquired information as a vital cog in our control system algorithm, thereby improving the intelligence of the home. We created a system capable of running with minimal human supervision, an attribute which makes our system an asset for senior care scenarios. Our system also caters to the safety of the home owner.

Partially connected ANN with Evolvable Topology (PANNET) is a non-fully connected recurrent neural network with proper number of context nodes. The structure of the network along with connection weights are determined through the evolutionary process of a customized genetic algorithm. In this paper, we develop an evolutionary bilevel optimization procedure for tuning the hyper-parameters of PANNET. In the upper level, an evolutionary algorithm optimizes the hyper-parameters, while the customized genetic algorithm is training the PANNET in the lower level optimization. Since executing the lower level optimization for each candidate hyper-parameters requires a high computational cost, fitness function approximation is performed using a regression model based on the Random Forest method. The proposed procedure provides more flexibility on choosing the hyper-parameters, and generate a smaller network with more accuracy in prediction.

The capability to estimate driver's intention leads to the development of advanced driver assistance systems that can assist the drivers in complex situations. Developing precise driver behavior models near intersections can considerably reduce the number of accidents at road intersections. In this study, the problem of driver behavior modeling near a road intersection is investigated using support vector machines (SVMs) based on the hybrid-state system (HSS) framework. In the HSS framework, the decisions of the driver are represented as a discrete-state system and the vehicle dynamics are represented as a continuous-state system. The proposed modeling technique utilizes the continuous observations from the vehicle and estimates the driver's intention at each time step using a multi-class SVM approach. Statistical methods are used to extract features from continuous observations. This allows for the use of history in estimating the current state. The developed algorithm is trained and tested successfully using naturalistic driving data collected from a sensor-equipped vehicle operated in the streets of Columbus, OH and provided by the Ohio State University. The proposed framework shows a promising accuracy of above 97% in estimating the driver's intention when approaching an intersection.

Reverse Engineering of Gene Regulatory Networks (GRN), i.e. Finding appropriate mathematical models to understand complex cellular systems, can be used in disease diagnosis, treatment, and drug design. There are fundamental gaps in the construction of GRN with regard to modeling of hidden/delayed interactions. Addressing these deficiencies is critical to understanding complex intracellular processes and enabling full use of the vast and ever-growing amount of available genomic data. Current modeling strategies either ignore or oversimplify time delays resulted from transcription and translation processes during gene expression. In addition, many research works do not account hidden variables such as transcription factors, repressors, small metabolites, DNA, microRNA species that regulate themselves and other genes but are not readily detectable on micro array experiments. To capture the effect of these parameters, in this paper, we utilize our developed Partially Connected Artificial Neural Networks with Evolvable Topology (PANNET) to find a more comprehensive model of GRN by considering the effects of unknown hidden variables and different time delays. This method is innovative, since the structure of the network has memory and internal states, which can model the unknown hidden variables and time delays. We furthermore use a new evolutionary optimization based on variable-length Genetic Algorithm (GA) to find a sparse structure of PANNET to predict the gene expression levels accurately. Finally we demonstrate the capability of PANNET in constructing GRN, including the effect of different delays and unknown hidden variables through modeling of E. Coli SOS inducible DNA repair system.

This paper presents a Radio Frequency based system for handling long-term drift of pose estimates for a robot performing odometer-based navigation in an indoor environment. The indoor environment is augmented with RFID tags and their associated door-markers to form a partially structured environment. To enhance the performance of the odometry, we adopted a Least Squares calibration approach to mitigate the effect of the systematic errors. The residual errors, mainly non-systematic, are handled by intermittent resetting of the robot's pose based on the global positioning references designed with the RFID tags and their associated door-markers. The results reveal that the long-term confidence in the estimated position improves about six times with this approach.

Computational prediction of protein function is an important field in functional genomics. Gene function prediction is a Hierarchical Multi Label Classification (HMC) problem where each gene can belong to more than one functional class simultaneously, while classes are structured in the form of hierarchy. HMC is becoming a necessity in many domains of applications as well. Crowding niching-Adaptive mutation (CAM) is a new proposed method for solving Hierarchical multi-label gene function prediction problem. The classification in CAM-HMC is structured in three different phases. In the first two phases, a sequential procedure is performed. In the first phase, a full cyclic evolutionary crowding algorithm based on new definition of distance between two individuals, and adaptive mutation is applied in order to find classification rules. In the second phase, all the examples that are covered by these rules are removed from the training data. This sequential procedure is repeated until most of the training examples are covered by CAM-HMC rules. In the third phase, consequent generation is determined to show the probability of coverage of each rule for each hierarchical class. Finally, this ratio is applied to classify testing data. Efficiency of this algorithm is displayed by comparing this algorithm with HMC-GA using Precision-Recall curves for three numerical datasets related to protein functions of the Saccharomyces Cerevisiae organism.

In this paper, a semi-supervised technique based on support vector machine (SVM) for image classification and a Locality Sensitive Hashing (LSH) based searching algorithm to search for similarity of satellite imagery is presented. Given a query image, the goal is to retrieve matching images in the database based on the shape features extracted from satellite imagery data. The experimental results demonstrate superior results based on shape features which provide a better classification accuracy using both support vector machine and the semi-supervised hashing search methods.

This paper develops an effective algorithm for simultaneous robust adaptive control (filtering) and unknown parameter estimation. The need for adaptive control of poorly known systems is recognized in a diverse range of applications from active noise cancellation and structural vibration isolation to temperature control in process industry, to equalization in communication systems. Due to the nonlinear dependency of the problem on the unknown parameters, an exact solution to this problem is not known to date. The approach discussed uses a robust adaptive filtering/control algorithm to fulfil the control objectives while a structured search algorithm (evolutionary algorithm) identifies the unknown (possibly changing) parameters. This approach overcomes unnecessary approximations that are commonly adopted in the existing solutions by casting the nonlinear parameter estimation problem as an optimization problem. Furthermore, the evolutionary algorithm formulation provides the framework to include the robustness of parameter estimates as an explicit objective in the optimization problem. Robust parameter estimates are of added significance when they are used in a feed-forward control path. Simulation results are presented to demonstrate the feasibility, performance, and the main features of the proposed approach

We sketch a framework for learning structured coordinated behavior, specifically the tactical behavior of experimental unmanned vehicles (XUVs). We conceptualize an XUV unit as a multiagent system (MAS) on which we impose a command structure to yield a holarchy, a hierarchy of holons, where a holon is both a whole and a part. The formalism used is a conservative extension of statecharts, called a parts/whole statechart, which introduces a coordinating whole as a concurrent component on a par with the coordinated parts; wholes are related to common knowledge. We use X-classifier systems (XCSs), where learning acquires a population of weighted condition-action classifier rules that direct behavior. Environmental rewards modify classifier strength, and a genetic algorithm (GA) modifies the classifier population. Exploiting statechart semantics, we translate statechart transitions into classifiers and define data structures that interact with an XCS. Difficulties arise in learning wholes and where the GA makes structural changes.

We address the formal representation of military tactical behaviors, in particular, the coordinated movements of experimental unmanned vehicles (XUVs). A unit of cooperating XUVs is considered a multiagent system on which the hierarchical structure of military command and control is imposed to yield a holarchy, a hierarchy of holons, where a holon is both a whole and a part. To model holarchical units of XUVs, we use a statechart extension called a parts/whole statechart, which introduces a coordinating whole as a concurrent component on a par with the coordinated parts. Implementing a part amounts to endowing an individual with the capacity for certain public behavior, while implementing a whole amounts to endowing the subsumed individuals with common knowledge hence the capacity for certain social behavior.

This paper applies the biofunctional model of human learning to the implementation of a learning machine that is effective in navigating complex environments. The target model is rule-based and is highly flexible in establishing the relation between any state-action pair. The learning machine is designed using X classifier systems and a fuzzy logic controller (FLC). A learning machine is built in simulation that closely approximates the learning characteristics of the human brain as described by the theory of biofunctional cognition. The methodology is tested with experiments using both single and multiple agents. We also investigated the effectiveness of biofunctionality using competitive and cooperative modes. Furthermore, we studied the robustness of our approach. Our results show that the integration of a FLC and an X classifier system, realizing a biofunctional model, provides a methodology for constructing very effective learning machines.

The first generation of learning scientists were behavioral psychologists, who pioneered the field in the first half of the 20th Century. Today, behavioral learning theory continues to enjoy a strong following even though its mainstream position was taken up by the second generation of learning researchers, namely cognitive psychologists. During the cognitive era, interest in the scientific study of learning spread to engineering and computer science with an almost exclusive focus on knowledge acquisition as the symbolic internalization of soft representations. Learning beyond soft knowledge, qualitative learning, and learning as multiple-source organizing were left up to the third generation of learning scientists - namely, biofunctionalists - who crossed the threshold of the 21st Century in dim popularity, mainly because of their counterintuitive definition of learning as whole theme reorganization of the learner's own intuitive knowledge base. This paper examines the three pioneering shifts in the science of learning with a focus on the nature of their landmark transitions.

We designed a feedforward control law that greatly decreases the load sway of a shipboard crane due to ship rolling. This feedforward control uses measurements of ship rolling angle at each instant. At different operating points the optimal feedforward gain changes while is numerically computable. Here, we propose to use a genetic algorithm (GA) based approach to optimize the mapping of feedforward gain in four dimensional space. The process is based on the numerical calculation of the optimal feedforward gain for any rolling angle (/spl rho/), length of the rope (L), and luffing angle (/spl delta//sub 0/). The optimal gain is calculated for a group of points in the working space and then fit a function of order n to these points in a four dimensional space. Our choice for this problem includes real value GA with a combination of different crossover methods. The cost function is the sum of squared errors at selected points and we aim to minimize it. Since moving the load to another location also changes the optimal gain, the new improved gain scheduling further reduces the swinging within the whole working space. GA is a directed search method and is capable of searching for variables of functions with any desired structure. The major advantages of using GA for function mappings is that the function does not have to be linear or in any specific form.

A multilayered multifuzzy logic controllers (MLMFLC) scheme is introduced to navigate the Khepera miniature mobile robot through obstacles and narrow opening hallways. Robots eight infrared proximity sensors are divided into three groups (left, right, and back sensors) and are fed to three fuzzy inference systems (FIS) in the first layer. The output of the first FIS group in the first layer is a representation of the robot's immediate surrounding obstacles. Three outputs of the three FIS in the first layer are the inputs of the second layer's FIS. Eventually, the output of the second layer FIS operates two step-motors according to position of the obstacles surrounding the robot. A total of nineteen rules have been employed for all FIS blocks. A real environment with obstacles and a dead-end trap has been used and experimental results on a real mobile robot have been demonstrated. By applying the controller, the robot moves smoothly in the experimental environment without any collision. Since the robot adjusts its speed in response to the environment, the overall speed is desirable. The design process and experimental set up is explained in detail.

Building a knowledge base for an intelligent system is the main goal in the development of any learning machine. Our daily lives and experiences suggest that human-like learning systems are better suited for functioning in hard-to-navigate environments because of their high degree of flexibility. This paper applies the biofunctional model of human learning to the design and implementation of a learning machine that is effective in navigating complex environments and relatively easy to design using classifier systems. We portray in the case study how a fuzzy logic controller links the system to the biofunctional model. It also makes vast improvements to the learning rate and the overall efficiency of the whole system.

Crane control is a difficult problem for conventional control methods because of the highly nonlinear equations that must be satisfied. Usually the necessary conditions for solving an optimal control problem require finding the initial co-state vector. In this paper real-coded genetic algorithms are used to find the desired initial value of the costates of the system with no constraints. In our genetic representation, each chromosome represents a set of co-states and each gene (co-state) has an associated cost based on its ability to move the system to desired state after a given amount of time. The objective is to evolve a minimum cost co-state. Our results for this unconstrained crane problem are quite encouraging.

Hybrid fuzzy-PID (HFPID) control, the cascade of a fuzzy and a conventional PID controller with the fuzzy controller acting as the designer of the PID controller, is applied to the control of the longitudinal axis of a VSTOL aircraft. The control of the longitudinal axis is a nonlinear problem complicated by the transition from hover to normal flight required by a VSTOL aircraft. This transition region may introduce singular matrices not usually encountered in a normal flight regime, hence requiring existing techniques to be compensated. A HFPID controller is designed and simulation results are presented.

We combine the concept of evolutionary search with the systematic search concepts of arc revision and hill climbing to form a hybrid system that quickly finds solutions to static and dynamic constraint satisfaction problems (CSPs). Furthermore, we present the results of two experiments. In the first experiment, we show that our evolutionary hybrid outperforms a well-known hill climber, the iterative descent method (IDM), on a test suite of 750 randomly generated static CSPs. These results show the existence of a "mushy region" which contains a phase transition between CSPs that are based on constraint networks that have one or more solutions and those based on networks that have no solution. In the second experiment, we use a test suite of 250 additional randomly generated CSPs to compare two approaches for solving CSPs. In the first method, all the constraints of a CSP are known by the hybrid at run-time. We refer to this method as the static method for solving CSPs. In the second method, only half of the constraints of a CSPs are known at run-time. Each time that our hybrid system discovers a solution that satisfies all of the constraints of the current network, one additional constraint is added. This process of incrementally adding constraints is continued until all the constraints of a CSP are known by the algorithm or until the maximum number of individuals has been created. We refer to this second method as the dynamic method for solving CSPs. Our results show hybrid evolutionary search performs exceptionally well in the presence of dynamic (incremental) constraints, then also illuminate a potential hazard with solving dynamic CSPs.

This paper investigates the role of hierarchy in the systematic approach to the design of fuzzy logic controllers (FLC's). The key concept here is that the implementation of fuzzy engines at higher levels of the control hierarchy (where more reasoning is involved) yields more versatile fuzzy controllers with generally fewer control rules. At the same time, the structured nature of a hierarchical approach considerably simplifies the design procedure.

This paper examines the applicability of genetic algorithms (GA's) in the simultaneous design of membership functions and rule sets for fuzzy logic controllers. Previous work using genetic algorithms has focused on the development of rule sets or high performance membership functions; however, the interdependence between these two components suggests a simultaneous design procedure would be a more appropriate methodology. When GA's have been used to develop both, it has been done serially, e.g., design the membership functions and then use them in the design of the rule set. This, however, means that the membership functions were optimized for the initial rule set and not the rule set designed subsequently. GA's are fully capable of creating complete fuzzy controllers given the equations of motion of the system, eliminating the need for human input in the design loop. This new method has been applied to two problems, a cart controller and a truck controller. Beyond the development of these controllers, we also examine the design of a robust controller for the cart problem and its ability to overcome faulty rules.>