Large language models (LLMs) present new opportunities for creating pedagogical agents that engage in meaningful dialogue to support student learning. However, current LLM systems used in classrooms often lack the solid theoretical foundations found in earlier intelligent tutoring systems. To bridge this gap, we propose a framework that combines Evidence-Centered Design with Social Cognitive Theory and Zone of Proximal Development for adaptive scaffolding in LLM-based agents focused on STEM+C learning. We instantiate this framework with Inquizzitor, an LLM-based formative assessment agent that integrates human-AI hybrid intelligence and provides feedback grounded in cognitive science principles. Our findings show that Inquizzitor delivers high-quality assessment and interaction aligned with core learning theories, offering effective guidance that students value. This research demonstrates the potential for theory-driven LLM integration in education, highlighting the ability of these systems to provide adaptive and principled instruction.

This paper presents an approach that uses Large Language Models (LLMs) as late-fusion interpreters to synthesize multimodal signals from embodied classroom activities and infer students’ metacognitive behaviors. Our multimodal pipeline analyzes students’ movements, gaze, gestures, and speech within a mixed-reality simulation displayed on a classroom screen to support enactment and learning. Vision- and speech-derived features are fused at the interpretive layer via zero-shot prompting, self-consistency reasoning, and targeted prompt engineering to derive planning, enacting, monitoring, reflecting, and interacting behaviors. We investigate whether LLMs can reliably integrate modality-specific analytics to produce accurate behavioral labeling and whether an LLM-as-a-Judge can validate them at scale. To address scalability and reduce human burden, we introduce an automated evaluation protocol employing LLM-as-a-Judge to assess classification quality, enabling rapid, iterative benchmarking of model variants and prompt strategies. Using a balanced corpus of human-validated segments and perturbed controls, we compare text-only language models (e.g., GPT-5) with visual–language models (e.g., Qwen2.5-VL) that incorporate direct visual processing. Results indicate late-fusion, text-based LLMs can outperform VLMs on behavior judgment without raw video, and precision- or recall-oriented prompts adjust decision boundaries for subtle or brief segments. These findings position LLMs as effective late-fusion mechanisms for multimodal learning analytics and demonstrate the viability of LLM-as-a-Judge for scalable, human-in-the-loop evaluation.

Problem-based learning (PBL) environments increasingly embed LLM-based conversational agents (CAs) to scaffold inquiry, yet little is known about how learners actually respond to these agents in authentic classroom settings. Learning analytics offers powerful opportunities to capture and interpret how students engage with these agents, enabling deeper understanding of their inquiry processes and informing more adaptive instructional support in PBL settings. In this paper, we examine students’ interactions with three types of LLM-powered CAs — Content Knowledge, Argument Feedback, Argument Evaluation — designed to provide distinct forms of inquiry support within a narrative-centered learning environment. Using Pedaste et al.’s inquiry cycle as a lens, we used contextualized log data from 15 student groups to analyze how these agents shaped inquiry via sequence analysis of students’ coded actions. Our results revealed distinct trajectories of agent episodes and suggest LLM-powered CAs can play complementary pedagogical roles — supporting information seeking, guiding revision, and prompting reflection — but may also channel inquiry in ways that constrain exploration. We discuss the implications of using learning analytics to design adaptive scaffolds and using contextualized log analysis to capture how learners navigate inquiry with AI support in authentic classroom settings.

The field of learning analytics has made notable strides in automating the detection of complex learning processes in multimodal data. However, most advancements have focused on individualized problem-solving instead of collaborative, open-ended problem-solving, which may offer both affordances (richer data) and challenges (low cohesion) to behavioral prediction. Here, we extend predictive models to automatically detect socially shared regulation of learning (SSRL) behaviors in collaborative computational modeling environments using embedding-based approaches. We leverage large language models (LLMs) as summarization tools to generate task-aware representations of student dialogue aligned with system logs. These summaries, combined with text-only embeddings, context-enriched embeddings, and log-derived features, were used to train predictive models. Results show that text-only embeddings often achieve stronger performance in detecting SSRL behaviors related to enactment or group dynamics (e.g., off-task behavior or requesting assistance). In contrast, contextual and multimodal features provide complementary benefits for constructs such as planning and reflection. Overall, our findings highlight the promise of embedding-based models for extending learning analytics by enabling scalable detection of SSRL behaviors, ultimately supporting real-time feedback and adaptive scaffolding in collaborative learning environments that teachers value.

Wearable cameras provide an efficient and convenient way to record our lives, supporting real-time documentation and analysis across various domains. Recent research has explored diverse methods for temporal segmentation, which aim to transform unstructured video data into structured events. This transformation facilitates deeper video understanding, optimizes computational resources, and improves the accessibility and interpretability of video content for both machines and humans. However, unlike conventional videos, wearable camera recordings present unique challenges. These include highly unstable camera perspectives, diverse activities across various environments, and flexible duration. As a result, traditional temporal segmentation methods often fail to return effective results. This paper introduces SmartSeg, an unsupervised, non-parametric approach for segmenting wearable camera videos without labeled data. By capturing the fundamental meanings of the video, SmartSeg aggregates the video through the Temporal Self-Similarity Metric encoder and groups sequences of frames into coherent events through clustering techniques. We evaluated SmartSeg on three diverse datasets. We achieved a 50% increase in Mean-over-Frames(MoF) compared to the state-of-the-art on one egocentric dataset. We conducted a real-world case study on nursing simulations, demonstrating SmartSeg’s ability to effectively segment complex, noisy interactions with diverse activity transitions. The results highlight SmartSeg’s robustness in handling long, unstructured, and visually challenging wearable camera videos, establishing it as a promising tool for real-world video temporal segmentation tasks.

Fault detection and isolation are becoming increasingly important as modern systems become more complex. To encourage the development of new fault detection solutions that can operate with limited noisy data and an incomplete mathematical model, the DX 2025 LiU-ICE competition for diagnosis of the air path of an internal combustion engine was introduced. In this paper, we present our winning solution to this competition. Our fault detection architecture starts with a semi-supervised Transformer Autoencoder trained to reconstruct nominal data. Detected faults are then passed through a rule-based fault persistence filter that aims to suppress false positives. Once a fault is detected, we use four neural networks trained to estimate features determined from structural analysis of a partial system model. The residuals of these networks are fed to a supervised fault classification network that estimates the fault probabilities. With this architecture, we achieved an 87% detection rate with a 0% false alarm rate on the provided competition data. Additionally, our isolation architecture assigned the correct fault 73.8% probabilty on average. On unseen competition data from a new driving cycle, we achieved a 100% detection rate and assigned the correct fault 66.2% probability on average. On the other hand, the Transformer Autoencoder failed to transfer to the new driving conditions, causing many false alarms. We discuss ways future work can reduce this.

Accurate estimation of the remaining useful life (RUL) of industrial systems is a critical component of predictive maintenance strategies. This work presents a data-driven method for RUL prediction that also quantifies uncertainty, drawing inspiration from model-based particle filtering techniques. Instead of simulating system state transitions, we model degradation as a stochastic process governed by performance metrics and use a Bayesian particle filtering framework to infer its underlying parameters. Our approach bypasses traditional state-space modeling by directly estimating the end-of-life distribution from observed performance data. Key characteristics of the filter, such as propagation noise and observation correction strength, are adapted over time based on current observations and past predictive performance, enabling better capture of future uncertainty. We evaluate the proposed method using an unmanned aerial vehicle simulation dataset developed for system-level prognostics research, which includes high-fidelity degradation signals and ground-truth system performance metrics for validating predictive accuracy.

Ensuring flight safety for small unmanned aerial systems (sUAS) requires continuous in-flight monitoring and decision-making, as unexpected events can alter power consumption and deplete battery energy faster than anticipated. Such events may result in insufficient battery capacity to complete a mission, thereby compromising flight safety. In this paper, we present an online feasibility assessment and contingency management framework that continuously monitors the aircraft’s battery state and the energy required to complete the flight in real-time, which enables informed decision-making to enhance flight safety. The framework consists of two main components: power consumption prediction and battery voltage trajectory prediction. The power consumption prediction is conducted using a model that is based on momentum theory, while the voltage trajectory prediction is performed using a Neural Ordinary Differential Equation (Neural ODE)-based data-driven model. By integrating these two components, the framework evaluates the feasibility of a flight mission in real time and determines whether to proceed with the mission or initiate rerouting. We evaluate the framework’s performance in a drone delivery scenario in the Dallas–Fort Worth (DFW) area, where the aircraft encounters an unexpected energy depletion event mid-flight. The proposed framework is tasked with assessing the feasibility of completing the mission and, if necessary, rerouting the aircraft for an emergency landing. The results demonstrate that the framework accurately and efficiently detects energy insufficiencies in real-time and re-routes the aircraft to a [3] predefined emergency landing site.

Offline reinforcement learning (RL) enables learning decision-making policies directly from historical data, which is advantageous for safety-critical domains like traffic control. However, existing offline RL benchmarks in transportation systems typically rely on simulated data, which may not fully capture the complexities of real-world environments. In this paper, we introduce the first offline RL benchmark for variable speed limit (VSL) control, built from approximately 100 million transitions of real-world interaction data collected from a field-deployed, multi-agent RL-based VSL system on a major freeway. We evaluate five state-of-the-art offline RL algorithms under multiple dataset conditions defined by varying sizes and action noise levels. Through traffic microsimulation experiments, we analyze algorithm performance and generalization, providing insights into the challenges and opportunities of offline RL for intelligent transportation systems. The dataset and benchmark are released at https://github.com/Lab-Work/i24-vsl-orl.

This paper explores the use of large language models (LLMs) to score and explain short-answer assessments in K-12 science. While existing methods can score more structured math and computer science assessments, they often do not provide explanations for the scores. Our study focuses on employing GPT-4 for automated assessment in middle school Earth Science, combining few-shot and active learning with chain-of-thought reasoning. Using a human-in-the-loop approach, we successfully score and provide meaningful explanations for formative assessment responses. A systematic analysis of our method's pros and cons sheds light on the potential for human-in-the-loop techniques to enhance automated grading for open-ended science assessments.

Computer-Based Open-Ended Learning Environments (OELEs) are designed to challenge learners to become proficient problem-solvers and develop the ability to independently solve complex problems. However, the traditional focus of OELE research has been on demonstrating overall learning gains, potentially overlooking students who struggle in these environments. To address this gap, we take a social justice-based approach by studying 99 sixth-grade students who participated in a week-long classroom study. We first assessed learning outcomes across all then identified 20 students who failed to do well. We qualitatively analyzed video recordings of their interactions with the OELE to understand why they struggled and to determine if interface issues inhibited their learning. Five themes emerged: (1) challenges in knowledge acquisition; (2) challenges in scaffolding learning; (3) disregarding system guidance, (4) not leveraging supporting tools; (5) and getting discouraged by incorrect answers. Based on our findings, we make design recommendations for OELEs to better support underserved learners, recognizing that failure is an important catalyst for motivating improvements in child-centered design.

Due to the safety risks and training sample inefficiency, it is often preferred to develop controllers in simulation. However, minor differences between the simulation and the real world can cause a significant sim-to-real gap. This gap can reduce the effectiveness of the developed controller. In this paper, we examine a case study of transferring an octorotor reinforcement learning controller from simulation to the real world. First, we quantify the effectiveness of the real-world transfer by examining safety metrics. We find that although there is a noticeable (around 100%) increase in deviation in real flights, this deviation may not be considered unsafe, as it will be within > 2m safety corridors. Then, we estimate the densities of the measurement distributions and compare the Jensen-Shannon divergences of simulated and real measurements. From this, we show that the vehicle’s orientation is significantly different between simulated and real flights. We attribute this to a different flight mode in real flights where the vehicle turns to face the next waypoint. We also find that the reinforcement learning controller actions appear to correctly counteract disturbance forces. Then, we analyze the errors of a measurement autoencoder and state transition model neural network applied to real data. We find that these models further reinforce the difference between the simulated and real attitude control, showing the errors directly on the flight paths. Finally, we discuss important lessons learned in the sim-to-real transfer of our controller.

As online learning tools become more widespread, understanding student behaviors through learning analytics is increasingly important. Traditional methods relying on system log data fall short of capturing the full range of cognitive strategies students use. To address this, we developed an in-depth post-assignment reflection dashboard that visualizes gaze data to aid students in reflecting on their learning behaviors. This dashboard was made possible by ETProWeb, a system that integrates high-fidelity eye-tracking directly into the browser, enabling real-time analysis of gaze data aligned with user interactions. ETProWeb leverages the browser's Document Object Model (DOM) to track areas of interest (AOIs) dynamically, overcoming issues related to multiple timelines and manual alignment. In a pilot study with 38 sixth-grade students, the dashboard received positive feedback, with 90% of students expressing interest in the eye-tracking technology for its ability to help them observe and reflect on their reading behaviors. This interest highlights the potential of eye-tracking as a valuable tool for enhancing students' self-awareness and engagement in online learning environments.

Embodied learning represents a natural and immersive approach to education, where the physical engagement of learners plays a critical role in how they perceive and internalize concepts. This allows students to actively embody and explore knowledge through interaction with their environment, significantly enhancing retention and understanding of complex subjects. However, researchers face significant challenges in exploring children's learning in these physically interactive spaces, particularly due to the complexity of tracking multiple students' movements and dynamic interactions in real-time. To address these challenges, this paper introduces a Double Diamond design thinking process for developing an AI-enhanced timeline aimed at assisting researchers in visualizing and analyzing interactions within embodied learning environments. We outline key considerations, challenges, and lessons learned in this user-centered design process. Our goal is to create a timeline that employs state-of-the-art AI techniques to help researchers interpret complex datasets, such as children's movements, gaze directions, and affective states during learning activities, thereby simplifying their tasks and augmenting the process of interaction analysis.

The incorporation of technology into primary and secondary education has facilitated the creation of curricula that utilize computational tools for problem-solving. In Open-Ended Learning Environments (OELEs), students participate in learning-by- modeling activities that enhance their understanding of (Science, technology, engineering, and mathematics) STEM and computational concepts. This research presents an innovative multimodal emotion recognition approach that analyzes facial expressions and speech data to identify pertinent learning-centered emotions, such as engagement, delight, confusion, frustration, and boredom. Utilizing sophisticated machine learning algorithms, including High-Speed Face Emotion Recognition (HSEmotion) model for visual data and wav2vec 2.0 for auditory data, our method is refined with a modality verification step and a fusion layer for accurate emotion classification. The multimodal technique significantly increases emotion detection accuracy, with an overall accuracy of 87%, and an Fl -score of 84%. The study also correlates these emotions with model building strategies in collaborative settings, with statistical analyses indicating distinct emotional patterns associated with effective and ineffective strategy use for tasks model construction and debugging tasks. These findings underscore the role of adaptive learning environments in fostering students' emotional and cognitive development.

While understanding Self-Regulated Learning (SRL) in Open-Ended Learning Environments (OELEs), it is crucial to examine the interplay between students’ cognitive processes and affective states, especially learning centered emotions like delight, engagement, boredom, frustration and confusion. These affective states are particularly challenging to detect using facial expressions in middle school students, primarily due to the scarcity of relevant databases. This study introduces a novel approach that utilizes the EmoNet framework, enhanced with self-attention networks, to detect and analyze learning-centered emotions. We investigated the emotional and cognitive dynamics of 41 middle school students within an OELE. Our findings demonstrate distinct emotional patterns that significantly correlate with students’ performance levels across various cognitive processes. By creating and analyzing a dataset from ten students, the proposed model achieved a test accuracy of 85%, indicating a substantial improvement over existing state-of-the-art models. These results lay the groundwork for future educational tools capable of adapting to a combination of students’ affective and cognitive states thus enhancing their overall learning experiences that influence their educational outcomes.

Numerous studies aim to enhance learning in digital environments through emotionally-sensitive interventions. The D’Mello and Graesser (2012) model of affect dynamics hypothesizes that when a learner encounters confusion, the degree to which it is prolonged (and transitions into frustration) or resolved, significantly affects their learning outcomes in digital environments. However, studies yield inconclusive results regarding relations between confusion, frustration, and learning. More research is needed to explore how confusion and frustration manifest during learning and its relation to outcomes. We go beyond past work looking at the rate, duration, and transitions of confusion and frustration by treating these affective states as non-linear dynamical systems consisting of expressive and behavioral components. We examined the frequency and recurrence of facial expressions associated with basic emotions (as automatically labeled by AffDex, a standard tool for analyzing emotions with video data) during confused and frustrated states (as automatically labeled with BROMP-based detectors applied to students’ interaction data). We compare these co-occurring patterns to learning outcomes (pre-tests, post-tests, and learning gains) within a digital learning environment, Betty’s Brain. Results showed that the frequency and recurrence rate of basic emotions expressed during confusion and frustration are complex and remain incompletely understood. Specifically, we show that confusion and frustration have different relationships with learning outcomes, depending on which basic emotion expressions they co-occur with. Implications of this study open avenues for better understanding these emotions as complex and non-linear dynamical systems, in the long-term enabling personalized feedback and emotional support within digital learning environments that enhance learning outcomes.

This study introduces a methodology to investigate students’ collaborative behaviors as they work in pairs to build computational models of scientific processes. We expand the Self-Regulated Learning (SRL) framework—specifically, Planning, Enacting, and Reflection—proposed in the literature, applying it to examine students’ collaborative problem-solving (CPS) behaviors in a computational modeling task. We analyze these behaviors by employing a Markov Chain (MC) modeling approach that scrutinizes students’ model construction and model debugging behaviors during CPS. This involves interpreting their actions in the system collected through computer logs and analyzing their conversations using a Large Language Model (LLM) as they progress through their modeling task in segments. Our analytical framework assesses the behaviors of high- and low-performing students by evaluating their proficiency in completing the specified computational model for a kinematics problem. We employ a mixed-methods approach, combining Markov Chain analysis of student problem-solving transitions with qualitative interpretations of their conversation segments. The results highlight distinct differences in behaviors between high- and low-performing groups, suggesting potential for developing adaptive scaffolds in future work to enhance support for students in collaborative problem-solving.

Early and accurate detection of anomalous events on the freeway, such as accidents, can improve emergency response and clearance. However, existing delays and mistakes from manual crash reporting records make it a difficult problem to solve. Current large-scale freeway traffic datasets are not designed for anomaly detection and ignore these challenges. In this paper, we introduce the first large-scale lane-level freeway traffic dataset for anomaly detection. Our dataset consists of a month of weekday radar detection sensor data collected in 4 lanes along an 18-mile stretch of Interstate 24 heading toward Nashville, TN, comprising over 3.7 million sensor measurements. We also collect official crash reports from the Tennessee Department of Transportation Traffic Management Center and manually label all other potential anomalies in the dataset. To show the potential for our dataset to be used in future machine learning and traffic research, we benchmark numerous deep learning anomaly detection models on our dataset. We find that unsupervised graph neural network autoencoders are a promising solution for this problem and that ignoring spatial relationships leads to decreased performance. We demonstrate that our methods can reduce reporting delays by over 10 minutes on average while detecting 75% of crashes. Our dataset and all preprocessing code needed to get started are publicly released at https://vu.edu/ft-aed/ to facilitate future research.

Multimodal data analysis provides profound insights into behaviors and interactions within various settings. However, the collection and analysis of this data in real-world scenarios are intricate and resource-intensive. To streamline these processes, we introduce ChimeraPy: an open-source, distributed streaming platform optimized for high-throughput data transfer across processing nodes within a computer cluster. The utility and performance of ChimeraPy are showcased through two benchmark applications, highlighting its capability to handle complex data environments.

In this paper, we leverage ideas from model-based control to address the sample efficiency problem of reinforcement learning (RL) algorithms. Accelerating learning is an active field of RL highly relevant in the context of time-varying systems. Traditional transfer learning methods propose to use prior knowledge of the system behavior to devise a gradual or immediate data-driven transformation of the control policy obtained through RL. Such transformation is usually computed by estimating the performance of previous control policies based on measurements recently collected from the system. However, such retrospective measures have debatable utility with no guarantees of positive transfer in most cases. Instead, we propose a model-based transformation, such that when actions from a control policy are applied to the target system, a positive transfer is achieved. The transformation can be used as an initialization for the reinforcement learning process to converge to a new optimum. We validate the performance of our approach through four benchmark examples. We demonstrate that our approach is more sample-efficient than fine-tuning with reinforcement learning alone and achieves comparable performance to linear-quadratic-regulators and model-predictive control when an accurate linear model is known in the three cases. If an accurate model is not known, we empirically show that the proposed approach still guarantees positive transfer with jump-start improvement.

The optimal energy performance of building operations is affected by component faults, which may go unnoticed for long periods of time. Significant energy savings can be achieved if faulty behaviors are detected and rectified in a timely manner. In this work, we adopt an unsupervised approach for anomaly detection that combines automatic data engineering using clustering methods with Long-Short Term Memory (LSTM)-based Autoencoders. First, data engineering is used to extract multiple operating modes from nominal data of building operations. Then, an LSTM-based Autoencoder is trained to capture the characteristics of non-linear and temporal dynamics for each operating mode. Finally, the ensemble of models can be used for anomaly detection after training has been completed. We benchmark variants of our approach against state-of-the-art Autoencoders for anomaly detection by using a recently developed experimental dataset provided by the ASHRAE Research Project RP-1312. Unsupervised anomaly detection is a challenging task due to the lack of faulty labels and the need to identify faults while avoiding false alarms. Our novel approach improves the average true positive rate for fault detection by 11.4% against a state-of-the-art plain LSTM Autoencoder while keeping the false alarm rate around 5 % without having to use labeled fault data.

Eye-tracking technology has expanded our ability to quantitatively measure human perception. This rich data source has been widely used to characterize human behavior and cognition. However, eye-tracking analysis has been limited in its applicability, as contextualizing gaze to environmental artifacts is non-trivial. Moreover, the temporal evolution of gaze behavior through open-ended environments where learners are alternating between tasks often remains unclear. In this paper, we propose temporal fully-weighted scanpath graphs as a novel representation of gaze behavior and combine it with a clustering scheme to obtain high-level gaze summaries that can be mapped to cognitive tasks via network metrics and cluster mean graphs. In a case study with nurse simulation-based team training, our approach was able to explain changes in gaze behavior with respect to key events during the simulation. By identifying cognitive tasks via gaze behavior, learners’ strategies can be evaluated to create online performance metrics and personalized feedback.

Keeping K-12 teachers engaged during students’ learning and problem solving in technology-enhanced, integrated problem-based learning (PBL) has been shown to support deeper student involvement, and, therefore, better success learning difficult science, computing, and engineering concepts and practices. However, students’ learning processes and corresponding difficulties are not easily noticed by teachers as students learn from these environments as processes are captured through mouse clicks, drag and drop actions, and other low-level activities. As such, teachers find it difficult to set up meaningful interactions with students while also maintaining the focus on student-centered learning. Little research has examined dashboard-supported responsive teaching practices for K-12 PBL. This study examined 8 teachers as they used a co-designed teacher dashboard to assess and respond to students’ learning and strategies during an integrated, PBL STEM curriculum. Teachers completed a series of 5 “planning period simulations” leveraging the dashboard and think-aloud protocols were implemented, supported by semi-structured interview questions, to enable the teachers to verbalize their thought and evaluation processes. Content analysis and epistemic network analysis were conducted to analyze the simulations. Understanding how teachers use dashboards to support evidence-based teaching practices during technology-enhanced curricula is critical for improving teacher support and preparation.

Accurate estimation of the battery state of charge (SOC) for unmanned aerial vehicles (UAV) in-flight monitoring is essential for the safety and survivability of the system. Successful physics-based models of the battery have been developed in the past, however, these models do not take into account the effects of mission profile and environmental conditions during flight on the battery power consumption. Recently, data-driven methods have become popular given their ease of use and scalability. Yet, most benchmarking experiments have been conducted on simulated battery datasets. In this work, we compare different data-driven models for battery SOC estimation of a hexacopter UAV system using real flight data. We analyze the importance of a number of flight variables under different environmental conditions to determine the factors that affect battery SOC over the course of the flight. Our experiments demonstrate that additional flight variables are necessary to create an accurate SOC estimation model through data-driven methods.

Variable speed limit (VSL) control has emerged as a promising traffic management strategy for enhancing safety and mobility. In this study, we introduce a multi-agent reinforcement learning framework for implementing a large-scale VSL system to address recurring congestion in transportation corridors. The VSL control problem is modeled as a Markov game, using only data widely available on freeways. By employing parameter sharing among all VSL agents, the proposed algorithm can efficiently scale to cover extensive corridors. The agents are trained using a reward structure that incorporates adaptability, safety, mobility, and penalty terms; enabling agents to learn a coordinated policy that effectively reduces spatial speed variations while minimizing the impact on mobility. Our findings reveal that the proposed algorithm leads to a significant reduction in speed variation, which holds the potential to reduce incidents. Furthermore, the proposed approach performs satisfactorily under varying traffic demand and compliance rates.

In this work, we address the problem of solving complex collaborative robotic tasks subject to multiple varying parameters. Our approach combines simultaneous policy blending with system identification to create generalized policies that are robust to changes in system parameters. We employ a blending network whose state space relies solely on parameter estimates from a system identification technique. As a result, this blending network learns how to handle parameter changes instead of trying to learn how to solve the task for a generalized parameter set simultaneously. We demonstrate our scheme's ability on a collaborative robot and human itching task in which the human has motor impairments. We then showcase our approach's efficiency with a variety of system identification techniques when compared to standard domain randomization. The code is available on Luke Bhan's Github.

Students in computerized learning environments often direct their own learning processes, which requires metacognitive awareness of what should be learned next. We investigated a novel method of measuring verbalized metacognition by applying natural language processing (NLP) to transcripts of interviews conducted in a classroom with 99 middle school students who were using a computerized learning environment. We iteratively adapted the NLP method for the linguistic characteristics of these interviews, then applied it to study three research questions regarding the relationships between verbalized metacognition and measures of 1) learning, 2) confusion, and 3) metacognitive problem-solving strategies. Verbalized metacognition was not directly related to learning, but was related to confusion and metacognitive problem-solving strategies. Results also suggested that interviews themselves may improve learning by encouraging metacognition. We discuss implications for designing computerized environments that support self-regulated learning through metacognition.

Confusion has been shown to be prevalent during complex learning and has mixed effects on learning. Whether confusion facilitates or hampers learning may depend on whether it is resolved or not. Confusion resolution, behind which is the resolution of cognitive disequilibrium, requires learners to possess some skills, but it is unclear what these skills are. One possibility may be metacognitive strategies (MS), strategies for regulating cognition. This study examined the relationship between confusion and actions related to MS in Betty's Brain, a computer-based learning environment. The results revealed that MS behavior differed during and outside confusion. However, confusion resolution was not related to MS behavior, and MS did not moderate the effect of confusion on learning.

This paper presents a study using a robotics platform for teaching computing and cybersecurity concepts to high school students. 38 students attended a week-long camp, starting with projects such as a simulation-only game and a simple autonomous driving program for the robots in order to learn and apply computational thinking (CT) and networking skills. They were then assigned a series of challenges that required developing progressively more advanced cybersecurity measures to protect their robots. This culminated in a final challenge that required implementing defensive measures such as encryption, secure key exchange and sequence numbers. We used an evidence-centered design framework to construct rubrics for grading student work. The pre- and post-test results show that the interventions helped students learn cybersecurity and CT concepts, but they had difficulties with networking concepts. These results correlate with scores from the game and the final challenge. Overall, surveys show that the competition-based robotics learning framework engaged students and supported their overall learning, but our intervention needs to be modified to help students learn networking concepts

Education research has explored the role of students' affective states in learning, but some evidence suggests that existing models may not fully capture the meaning or frequency of how students transition between different states. In this study we examine the patterns of educationally-relevant affective states within the context of Betty's Brain, an open-ended, computer-based learning system used to teach complex scientific processes. We examine three types of affective transitions based on similarity with the theorized D'Mello and Graesser model, transition between two affective states, and the sustained instances of certain states. We correlate of the frequency of these patterns with learning outcomes and our findings suggest that boredom is a powerful indicator of students' knowledge, but not necessarily indicative of learning. We discuss our findings within the context of both research and theory on affect dynamics and the implications for pedagogical and system design.

C2STEM is a web-based learning environment founded on a novel paradigm that combines block-structured, visual programming with the concept of domain specific modeling languages (DSMLs) to promote the synergistic learning of discipline-specific and computational thinking (CT) concepts and practices. Our design-based, collaborative learning environment aims to provide students in K-12 classrooms with immersive experiences in CT through computational modeling in realistic scenarios (e.g., building models of scientific phenomena). The goal is to increase student engagement and include inclusive opportunities for developing key computational skills needed for the 21st century workforce. Research implementations that include a semester-long high school physics classroom study have demonstrated the effectiveness of our approach in supporting synergistic learning of STEM and CS/CT concepts and practices, especially when compared to a traditional classroom approach. This technology demonstration will showcase our CS+X (X = physics, marine biology, or earth science) learning environment and associated curricula. Participants can engage in our design process and learn how to develop curricular modules that cover STEM and CS/CT concepts and practices. Our work is supported by an NSF STEM+C grant and involves a multi-institutional team comprising Vanderbilt University, SRI International, Looking Glass Ventures, Stanford University, Salem State University, and ETR. More information, including example computational modeling tasks, can be found at C2STEM.org.

The paper presents RoboScape, a collaborative, networked robotics environment that makes key ideas in computer science accessible to groups of learners in informal learning spaces and K-12 classrooms. RoboScape is built on top of NetsBlox, an open-source, networked, visual programming environment based on Snap! that is specifically designed to introduce students to distributed computation and computer networking. RoboScape provides a twist on the state of the art of robotics learning platforms. First, a user's program controlling the robot runs in the browser and not on the robot. There is no need to download the program to the robot and hence, development and debugging become much easier. Second, the wireless communication between a student's program and the robot can be overheard by the programs of the other students. This makes cybersecurity an immediate need that students realize and can work to address. We have designed and delivered a cybersecurity summer camp to 24 students in grades between 7 and 12. The paper summarizes the technology behind RoboScape, the hands-on curriculum of the camp and the lessons learned.

This paper proposes an efficient online approach for reducing energy consumption in large buildings by combining data driven models with deep reinforcement learning techniques. We use data driven methods for modeling the heating and cooling energy consumption in the building. These models are integrated into a single "OpenAI Gym" class in Python to create the environment for studying building energy consumption as a function of control actions, such as setting the discharge temperature set points at different locations in the building. We discuss a policy gradient based actor-critic reinforcement learning approach (Q Actor-Critic) that learns the optimal policy by interacting with the above environment. The optimal policy acts as a controller for adjusting the discharge temperature set point of the dehumidified air in real time so that the total energy consumption can be reduced but the building conditions (temperature and humidity) remain comfortable. Preliminary results show that the method %is fast enough for online application and achieves an energy savings of 2 to 5%.

Modeling of HVAC components and energy flows for energy prediction purposes can be computationally expensive in large commercial buildings. More recently, the increased availability of building operational data has made it possible to develop data-driven methods for predicting and reducing energy use for these buildings. In this paper, we present such an approach, where we combine unsupervised and supervised learning algorithms to develop a robust method for energy reduction for large buildings operating under different environmental conditions. We compare our method against other energy prediction models that have been discussed in the literature using (1) a benchmark data set and (2) a real data set obtained from a building on the Vanderbilt University campus. A Stochastic Gradient Descent method is then applied to tune the controlled variable i.e., the AHU discharge temperature set point so that energy consumption is "minimized".

The relationship between learners' cognitive and affective states has become a topic of increased interest, especially because it is an important component of self-regulated learning (SRL) processes. This paper studies sixth grade students' SRL processes as they work in Betty's Brain, an agent-based open-ended learning environment (OELE). In this environment, students learn science topics by building causal models. Our analyses combine observational data on student affect to log files of students' interactions within the OELE. Preliminary analyses show that two relatively infrequent affective states, boredom and delight, show especially marked differences among high and low performing students. Further analysis shows that many of these differences occur after receiving feedback from the virtual agents in the Betty's Brain environment. We discuss the implications of these differences and how they can be used to construct adaptive personalized scaffolds.

Researchers have long recognized the importance of using technology to support students' collaboration in learning and problem solving tasks. Recently, there has been a lot of research in capturing and characterizing student discourse and how they regulate each other when they perform learning tasks in pairs or in small groups. In this paper, our goal is to dive a little deeper into how students collaborate, and the learning behaviors they exhibit when working in pairs on a learning by modeling task, while also teaching a virtual agent in the Betty's Brain system. We report the results of a quasi-experimental study, where students were divided into two groups: one group worked in pairs and the other group worked individually. The results illustrate that students in the collaborative group built more correct causal maps than those working individually, and their pre-post test results show significantly higher learning gains in the science content. A differential sequence mining algorithm applied to their action sequences captured in log files showed differences in the learning behaviors between the two groups. The differences imply that the collaborative groups were better at debugging their evolving causal maps than the students who worked individually.

This paper develops a structural diagnosis approach for fault detection and isolation in hybrid systems. Hybrid systems are characterized by continuous behaviors that are interspersed with discrete mode changes in the system, making the analysis of behaviors quite complex. In this paper, we address the mode detection problem in hybrid systems as the first step in diagnoser design. The proposed method uses analytic redundancy methods to detect the operating mode of the system even in the presence of system faults. We define hybrid minimal structurally overdetermined (HMSO) sets for hybrid systems. For residual generation, we develop the HMSO selection problem, formulated as a binary integer linear programming optimization problem to minimize the number of selected HMSOs and reduce online computational costs of the diagnosis algorithm. The proposed structural approach does not require preenumeration of all possible modes in the diagnoser design step. Therefore, our approach is feasible for hybrid systems with a large number of switching elements, implying that the system can have a large number of operating modes. The case study demonstrates the effectiveness of our approach. We discuss the results of our case study, and present directions for future work.

This paper combines a residual-based diagnosis approach and an unsupervised anomaly detection method to develop a hybrid methodology for monitoring smart buildings for which complete models are not available. The proposed method combines data mining approach and model-based diagnosis to update a diagnosis reference model and improve the overall diagnostics performance. To estimate the likelihood of each potential fault in complex systems like smart buildings, the dependencies between components and, there- fore, the monitors should be considered. In this work, a tree augmented naive Bayesian learning algorithm (TAN) is used for the classification. We demonstrate and validate the proposed approach using a data-set from an outdoor air unit (OAU) system in the Lentz public health center in Nashville.

Fault detection and isolation schemes are designed to detect the onset of adverse events during operations of complex systems, such as aircraft and industrial processes. The state-of-the-art fault diagnosis systems on aircraft combine an expert-created reference model of the associations between faults and symptoms, and a Naïve Bayes reasoner. For complex systems with many dependencies between components, the expert-generated reference models are often incomplete, which hinders timely and accurate fault diagnosis. Mining aircraft flight data is a promising approach to finding these missing relations between symptoms and data. However, mining algorithms generate a multitude of relations, and only a small subset of these relations may be useful for improving diagnoser performance. In this paper, we adopt a knowledge engineering approach that combines data mining methods with human expert input to update an existing reference model and improve the overall diagnostic performance. We discuss three case studies to demonstrate the effectiveness of this method.

Computational Thinking (CT) is considered a core competency in problem formulation and problem solving. We have developed the Computational Thinking using Simulation and Modeling (CTSiM) learning environment to help middle school students learn science and CT concepts simultaneously. In this paper, we present an approach that leverages multiple linked representations to help students learn by constructing and analyzing computational models of science topics. Results from a recent study show that students successfully use the linked representations to become better modelers and learners.

The detection and analysis of students’ domain-specific and metacognitive strategy use in Open-Ended Learning Environments (OELE) is a necessary step to support their learning and problem solving through contextualized scaffolding. We present an analysis of students’ performance from information captured in log files in UrbanSim, a turn-based simulation environment for counterinsurgency training. We illustrate the benefits of this approach within a task-model framework. Our overall goals are to implement a generalizable detection and adaptive scaffolding framework in an extended version of the GIFT tutoring system developed at ARL.

In this paper, we present early research evaluating the predictive power of a variety of temporal features across student subpopulations with distinctive behaviors at the beginning of the course. Initial results illustrate that these features predict important differences across the subpopulations and over time in the courses. Ultimately, these results have implications for effectively targeting adaptive scaffolding tailored to the particular intentions and goals of subpopulations in MOOCs.

The systems dynamics and control engineering (FDI) and the artificial intelligence diagnosis (DX) communities have developed complementary approaches that exploit structural relations in the system model to find efficient solutions for the residual generation and residual evaluation steps in fault detection and isolation in dynamic systems. This paper compares three different structural fault diagnosis techniques, two from the DX community and one from the FDI community. To simplify our comparison, we start with bond graphs as the common system modeling language and develop a graph-based framework using temporal causal graphs as the basis for analyzing the three fault isolation approaches. This framework allows for systematic comparison of the diagnosability properties of the three algorithms. The three-tank system is used as a running example to illustrate our concepts and algorithms.

Computational thinking (CT) draws on fundamental computer science concepts to formulate and solve problems, design systems, and understand human behavior. CT practices (e.g., problem representation, abstraction, decomposition, simulation, verification, and prediction) are also central to the development of expertise in a variety of STEM disciplines. Exploiting this synergy between CT and STEM disciplines, we have developed CTSiM, a cross-domain, scaffolded, visual-programming and agent-based learning environment for middle school science. We present and justify the CTSiM architecture and its implementation. To identify challenges and scaffolding needs in learning with CTSiM, we present a case study describing the challenges that a highand a low-achieving student faced while working on kinematics and ecology units using CTSiM. Decreases in the number of challenges for both students over sequences of related activities illustrate the combined effectiveness of our approach. Further, the specific challenges and scaffolds identified suggest the design of an adaptive scaffolding framework to help students develop a synergistic understanding of CT and science concepts.

graphs This paper discusses a set of semantic constraints that have to be applied for multi-domain modeling of complex, embedded systems. In particular, using the Hybrid Bond Graph (HBG) modeling language, we analyze issues that deal with consistent causality as-signments across model reconfigurations using hybrid switching junctions, and the complementarity of the electrical and mechanical domains by imposing addi-tional constraints in the modeling environment. A case study of a Reverse Osmosis system developed at NASA JSC illustrates the effectiveness of our approach.

This paper reports the results of an experiment that used qualitative and quantitative methods to investigate the effect of competition on students' use of game-based deliberate practice. We hypothesized that the results of the experiment would show that competition has a positive effect on performance outcomes, but it also increases students' tendency to game the system. The actual results of the experiment showed only very modest support for these hypotheses but have other implications for improving the design of educational games.

Metacognition and self-regulation are important components for developing effective learning in the classroom and beyond, but novice learners often lack these skills. Betty’s Brain, an open-ended computer-based learning environment, helps students develop metacognitive strategies as they learn science topics. In order to better understand and improve the effect of adaptive scaffolding on students’ cognitive and metacognitive skills, we investigate students’ activities in Betty’s Brain from a study comparing different forms of adaptive scaffolding. We measure students’ cognitive and metacognitive processes from students’ action sequences by (i) interpreting and characterizing behavior patterns using a cognitive/metacognitive model of the task, (ii) mapping students’ frequently observed cognitive and metacognitive process patterns back into their overall activity sequences and measuring their effectiveness, and (iii) employing a binning method with clustering and visualization techniques to characterize the temporal evolution of these processes. Our experimental studies illustrate that the effectiveness and temporal changes in students’ behaviors were generally consistent with the scaffolding provided, suggesting that these metacognitive strategies can be taught to middle school students in computer-based learning environments.

Open-ended learning environments (OELEs) provide students with opportunities to take part in authentic and complex problem-solving tasks. However, many students struggle to succeed in such complex learning endeavors. Without support, these students often use system tools incorrectly and adopt suboptimal learning strategies. However, providing adaptive support to students in OELEs poses significant challenges, and relatively few OELEs provide students with adaptive support. This paper presents the initial development of a systematic approach for interpreting and evaluating learner behaviors in OELEs called model-driven assessments, which uses a model of the cognitive and metacognitive processes important for completing the open-ended learning task. The model provides a means for both classifying and assessing students' learning behaviors while using the system. An evaluation of the analysis technique is presented in the context of Betty's Brain, an OELE designed to help middle school students learn about science.

The design of wireless sensor networks for detection applications is a challenging task. On one hand, classical work on decentralized detection does not consider practical wireless sensor networks. On the other hand, practical sensor network design approaches that treat the signal processing and communication aspects of the sensor network separately result in sub optimal detection performance because network resources are not allocated efficiently. In this work, we attempt to cross the gap between theoretical decentralized detection work and practical sensor network implementations. We consider a cross-layer approach, where the quality of information, channel state information, and residual energy information are included in the design process of tree-topology sensor networks. The design objective is to specify which sensors should contribute to a given detection task, and to calculate the relevant communication parameters. We compare two design schemes: (1) direct transmission, where raw data are transmitted to the fusion center without compression, and (2) in-network processing, where data is quantized before transmission. For both schemes, we design the optimal transmission control policy that coordinates the communication between sensor nodes and the fusion center. We show the performance improvement for the proposed design schemes over the classical decoupled and maximum throughput design approaches.

Computational Thinking (CT) defines a domain-general, analytic approach to problem solving that combines concepts fundamental to computing, with systematic representations for concepts and problem-solving approaches in scientific and mathematical domains. We exploit this trade-off between domain-specificity and domain-generality to develop CTSiM (Computational Thinking in Simulation and Modeling), a cross-domain, visual programming and agent-based learning environment for middle school science. CTSiM promotes inquiry learning by providing students with an environment for constructing computational models of scientific phenomena, executing their models using simulation tools, and conducting experiments to compare the simulation behavior generated by their models against that of an expert model. In a preliminary study, sixth-grade students used CTSiM to learn about distance-speed-time relations in a kinematics unit and then about the ecological process relations between fish, duckweed, and bacteria occurring in a fish tank system. Results show learning gains in both science units, but this required a set of scaffolds to help students learn in this environment.

Fault isolation and identification are necessary components for system reconfiguration and fault adaptive control in complex systems. However, accurate and timely on-line fault identification in nonlinear systems can be difficult and computationally expensive. In this paper, we improve the quantitative fault identification scheme in the TRANSCEND diagnosis approach. First, we propose to use possible conflicts (PCs) to find the set of minimally redundant subsystems that can be used for parameter estimation. Second, we introduce new algorithms for computing PCs from the temporal causal graph model used in TRANSCEND. Third, we use the minimal estimators to decompose the system model into smaller, independent subsystems for the parameter estimation task. We demonstrate the feasibility of this method by running experiments on a simulated model of the reverse osmosis subsystem of the advanced water recovery system developed at the NASA Johnson Space Center. Our results show a considerable reduction in parameter estimation time without loss of accuracy and robustness in the estimation.

This paper is focused on structural approaches to study diagnosability properties given a system model taking into account, both simultaneously or separately, integral and differential causal interpretations for differential constraints. We develop a model characterization and corresponding algorithms, for studying system diagnosability using a structural decomposition that avoids generating the full set of system analytical redundancy relations. Simultaneous application of integral and differential causal interpretations for differential constraints results in a mixed causality interpretation for the system. The added power of mixed causality is demonstrated using a Reverse Osmosis Subsystem from the Advanced Water Recovery System developed at the NASA Johnson Space Center. Finally, we summarize our work and provide a discussion of the advantages of mixed causality over just derivative or just integral causality.

This paper combines Multi-Agent based simulation with causal modeling and reasoning to help students learn about ecological processes. Eighth grade students who took part in the study showed highly significant pre to post test gains on learning domain content and causal reasoning ability. Moreover, students’ success in reasoning with a causal model of the ecosystem was strongly correlated with higher learning gains. This work provides the foundations for designing scaffolded multi-agent, simulation-based intelligent learning environments with modeling and reasoning tools to help students learn science topics.

The application of model-based diagnosis schemes to real systems introduces many significant challenges, such as building accurate system models for heterogeneous systems with complex behaviors, dealing with noisy measurements and disturbances, and producing valuable results in a timely manner with limited information and computational resources. The Advanced Diagnostics and Prognostics Testbed (ADAPT), which was deployed at the NASA Ames Research Center, is a representative spacecraft electrical power distribution system that embodies a number of these challenges. ADAPT contains a large number of interconnected components, and a set of circuit breakers and relays that enable a number of distinct power distribution configurations. The system includes electrical dc and ac loads, mechanical subsystems (such as motors), and fluid systems (such as pumps). The system components are susceptible to different types of faults, i.e., unexpected changes in parameter values, discrete faults in switching elements, and sensor faults. This paper presents Hybrid Transcend, which is a comprehensive model-based diagnosis scheme to address these challenges. The scheme uses the hybrid bond graph modeling language to systematically develop computational models and algorithms for hybrid state estimation, robust fault detection, and efficient fault isolation. The computational methods are implemented as a suite of software tools that enable diagnostic analysis and testing through simulation, diagnosability studies, and deployment on the experimental testbed. Simulation and experimental results demonstrate the effectiveness of the methodology.

The six papers in this special issue cover different approaches and different applications to model-based diagnostics.

This paper describes our initial efforts at implementing a new Choice-Adaptive Intelligent Learning Environment (CAILE) that combines multi-agent adaptive technologies and service architectures to provide a framework for designing extendible and reconfigurable learning environments. We describe the core components of the CAILE architecture, learning tasks that establish a situated context for learning, and a set of customizable agents that support student learning. We employ software engineering metrics to evaluate the system, and illustrate the reconfigurable and extensible properties of our design and implementation.

Sensor webs are often composed of servers connected to distributed real-time embedded (DRE) systems that operate in open environments where operating conditions, workload, resource availability, and connectivity cannot be accurately characterized a priori. The South East Alaska MOnitoring Network for Science, Telecommunications, Education, and Research (SEAMONSTER) project exhibits many common system management and dynamic operation challenges for effective, autonomous system adaptation in a representative sensor web. These challenges cover both field operation (e.g., power management through system sleep/wake cycles and reaction to local environmental changes) and server operation (e.g., system adaptation for new/modified goals, resource allocation for a changing set of applications, and configuration changes for fluctuating workload). This paper presents the results of integrating and applying quality-of-service (QoS)-enabled component middleware, dynamic resource management, and autonomous agent technologies to address these challenges in SEAMONSTER.

Wear and tear from sustained operations cause systems to degrade and develop faults. Online fault diagnosis schemes are necessary to ensure safe operation and avoid catastrophic situations, but centralized diagnosis approaches have large memory and communication requirements, scale poorly, and create single points of failure. To overcome these problems, we propose an online, distributed, model-based diagnosis scheme for isolating abrupt faults in large continuous systems. This paper presents two algorithms for designing the local diagnosers and analyzes their time and space complexity. The first algorithm assumes the subsystem structure is known and constructs a local diagnoser for each subsystem. The second algorithm creates the partition structure and local diagnosers simultaneously. We demonstrate the effectiveness of our approach by applying it to the Advanced Water Recovery System developed at the NASA Johnson Space Center.

Real-time and embedded systems have traditionally been designed for closed environments where operating conditions, input workloads, and resource availability are known a priori and are subject to little or no change at runtime. There is an increasing demand, however, for autonomous capabilities in open distributed real-time and embedded (DRE) systems that execute in environments where input workload and resource availability cannot be accurately characterized a priori. These systems can benefit from autonomic computing capabilities, such as self-(re)configuration and self-optimization, that enable autonomous adaptation under varying—even unpredictable—operational conditions. A challenging problem faced by researchers and developers in enabling autonomic computing capabilities to open DRE systems involves devising adaptive planning and resource management strategies that can meet mission objectives and end-to-end quality of service (QoS) requirements of applications. To address this challenge, this paper presents the Integrated Planning, Allocation, and Control (IPAC) framework, which provides decision-theoretic planning, dynamic resource allocation, and runtime system control to provide coordinated system adaptation and enable the autonomous operation of open DRE systems. This paper presents two contributions to research on autonomic computing for open DRE systems. First, we describe the design of IPAC and show how IPAC resolves the challenges associated with the autonomous operation of a representative open DRE system case study. Second, we empirically evaluate the planning and adaptive resource management capabilities of IPAC in the context of our case study. Our experimental results demonstrate that IPAC enables the autonomous operation of open DRE systems by performing adaptive planning and management of system resources.

This paper discusses the Bettypsilas Brain system and our ongoing work on developing a suite of tools that assist students and teachers in classroom learning in science domains. We describe the design and implementation of the system using a client/server architecture, and the initial responses of the teachers to the tools we have developed. Future enhancements and additions to the suite of tools are discussed.

Effective resource management for distributed real-time embedded (DRE) systems is hard due to their unique characteristics, including (1) constraints in multiple resources and (2) highly fluctuating resource availability and input workload. DRE systems can benefit from a middleware framework that enables adaptive resource management algorithms to ensure application QoS requirements are met. This paper identifies key challenges in designing and extending resource allocation algorithms for DRE systems. We present an empirical study of bin-packing algorithms enhanced to meet these challenges. Our analysis identifies input application patterns that help generate appropriate heuristics for using these algorithms effectively in DRE systems.

The two-party Diffie–Hellman (DH) key-exchanging technique is extended to generate (i) multiple two-party keys and (ii) one multi-party key. The participants in the former case exchange two public keys and generate 15 shared keys. Of these, 4 keys are called base keys, because they are used to generate the other 11 keys called extended keys. The main advantages are the reduction of the key exchange overhead, increase of additional protection to the keys and widening of applicability. In the latter case, an efficient contributory multi-party key-exchanging technique for a large static group is proposed. In this technique, a member who acts as a group controller forms two-party groups with other group members and generates a DH-style shared key per group. It then combines these keys into a single multi-party key and acts as a normal group member. The proposed technique has been compared with other multi-party key-generating techniques, and satisfactory results have been obtained.

Distributed real-time embedded (DRE) systems perform sequences of coordination and heterogeneous data manipulation tasks in dynamic environments to meet specified goals. Autonomous operation of DRE systems can benefit from the integrated operation of (1) a decision-theoretic spreading activation partial order planner (SA-POP) that combines task planning and scheduling in uncertain environments with (2) a resource allocation and control engine (RACE) middleware framework that integrates multiple resource management algorithms for (re)deploying and (re)configuring task sequence components in these systems. This paper demonstrates the effectiveness of SA-POP and RACE in managing and executing mission goals for a multisatellite application. Our results show that combining planning, scheduling and resource constraints dynamically is the key to implementing autonomy in DRE systems

Multirobot systems are being increasingly used for a variety of tasks in manufacturing, surveillance, and space exploration. These systems can degrade or develop faults during operation, and, therefore, require online diagnosis algorithms to ensure safe operation. Centralized approaches to online diagnosis of robot formations do not scale well for two primary reasons: 1) the computational complexity of the algorithm grows significantly with the number of robots, and 2) the individual robots must communicate a large number of measurements to a central diagnoser. To overcome these problems, we present a distributed, model-based, qualitative fault-diagnosis approach for formations of mobile robots. The approach is based on a bond-graph modeling framework that can deal with multiple sensor types and isolate process, sensor, and actuator faults. The diagnosis scheme employs relative measurement orderings to discriminate among faults by exploiting the temporal order of measurement deviations. This increases the discriminatory power of the measurement set and produces a more efficient fault-isolation algorithm. We describe a distributed diagnoser design algorithm applied to robot formations. Experimental results demonstrate the improvement in both the discriminatory power of the measurements produced by the relative measurement orderings, and the computational efficiency achieved by the distributed-diagnosis approach

Techniques for diagnosing faults in hybrid systems that combine digital (discrete) supervisory controllers with analog (continuous) plants need to be different from those used for discrete or continuous systems. This paper presents a methodology for online tracking and diagnosis of hybrid systems. We demonstrate the effectiveness of the approach with experiments conducted on the fuel-transfer system of fighter aircraft

Fault diagnosis of coupled mobile robots requires a large number of measurements to be communicated either between the robots or from the robots to a central diagnoser. As computational complexity increases with the number of measurements, centralized algorithms become inefficient. This paper presents a distributed approach for qualitative fault diagnosis of coupled mobile robots. The approach is based on a bond graph modeling framework which incorporates local and distributed control algorithms, multiple sensor types, and both actuator and sensor faults. Relative measurement orderings are introduced to discriminate faults by exploiting the temporal order of the measurement deviations. This increases the discriminatory power of a set of measurements and results in a more efficient qualitative diagnosis algorithm. Distributed diagnosers are designed and applied to coupled mobile robots. Experimental results for a system consisting of two robots pushing a box demonstrate the improvement in both discriminatory power of the measurements and efficiency of the distributed diagnosis approach

Embedded software and systems are at the intersection of electrical engineering, computer engineering, and computer science, with, increasing importance, in mechanical engineering. Despite the clear need for knowledge of systems modeling and analysis (covered in electrical and other engineering disciplines) and analysis of computational processes (covered in computer science), few academic programs have integrated the two disciplines into a cohesive program of study. This paper describes the efforts conducted at Vanderbilt University to establish a curriculum that addresses the needs of embedded software and systems. Given the compartmentalized nature of traditional engineering schools, where each discipline has an independent program of study, we have had to devise innovative ways to bring together the two disciplines. The paper also describes our current efforts in using learning technology to construct, manage, and deliver sophisticated computer-aided learning modules that can supplement the traditional course structure in the individual disciplines through out-of-class and in-class use.

Our group has been designing and implementing learning environments that promote deep understanding and transfer in complex domains. We have adopted the learning by teaching paradigm, and developed computer-based agents that students teach, and learn from this experience. The success of teachable agents has led us to develop a multiagent architecture that will be used to develop extended instructional systems based on gaming environments.

This paper describes the interface components for a system called Bettys Brain, an intelligent agent we have developed for studying the learning by teaching paradigm. Our previous studies have shown that students gain better understanding of domain knowledge when they prepare to teach others versus when they prepare to take an exam. This finding has motivated us to develop computer agents that students teach using concept map representations with a visual interface. Betty is intelligent not because she learns on her own, but because she can apply qualitative-reasoning techniques to answer questions that are directly related to what she has been taught through the concept map. We evaluate the agents interfaces in terms of how well they support learning activities, using examples of their use by fifth grade students in an extensive study that we performed in a Nashville public school. A critical analysis of the outcome of our studies has led us to propose the next generation interfaces in a multi-agent paradigm that should be more effective in promoting constructivist learning and self-regulation in the learning by teaching framework

Bettys Brain [1] is a learning-by-teaching environment where students "teach" Betty by constructing a concept map that models relations between domain concepts. The relations can be causal, hierarchical, and property links between the entities that represent the domain. The goal is for students to understand and then teach Betty about interdependence and balance among entities in a river ecosystem. As a part of the teaching process, students can query and quiz Betty to assess her understanding based on what she has been taught.Students can query Betty by asking her two types of questions: (i) "What happens to when increase/decrease?" and (ii) "Tell me about ". Betty answers questions by employing a qualitative reasoning mechanism, and explains her answers verbally and by using animation. Bettys explanations, and some feedback on the correctness of the answers should prompt the students to think more deeply about the domain processes, and motivate them to learn better before they teach her againStudents can request external feedback by asking Betty to take quizzes that are administered by a teacher agent. The teacher agent uses an overlay model to provide hints about concepts and links missing from the concept map. The hint levels start from general (e.g., suggesting that the student read a particular resource) to specific (e.g. indicating that a link is missing between two specific concepts. By seeing the quiz questions, students become aware of which concepts are important to model domain phenomena. The feedback from the teacher agent, points the students to understanding interrelationships among conceptsResults from our most recent study indicate that the query feature appeared to be effective in helping students develop an understanding of the interrelationships of living and non-living things in an ecosystem. The quiz feature is effective in helping students decide the important domain concepts and types of relationships to teach Betty. However, our observations of students during the study suggest that students using the quiz feature may have been overly focused on "getting the quiz questions correct" rather than "making sure that Betty (and themselves) understood the information.To help students focus more on learning, the next versions of Betty and teacher agent will be more interactive and metacognitive. In addition, the teacher will provide feedback that is related more to the global issues of balance and interdependence instead of individual links.At the demonstration, we will have a working version of our teachable agent system, Bettys Brain. We will demonstrate its user interfaces for the student to create and modify their concept maps, the query interface, and Bettys response to queries, and the teacher agent interface that responds when the students ask for help. In addition, the system also contains online resources that students can refer to when they are creating and updating their concept maps.

In this paper we propose a method to create discrete abstraction of state space behavior for continuous-time systems based on gradient analysis of the system dynamics. Then we describe how to use such a discrete model to design a supervisory controller for a given safety specification for the system. Finally we provide an entropy measure of nondeterminism, which can be used to evaluate the quality of the result discrete model as the degree of nondeterminism in that model.

Presents a similarity-based agglomerative clustering (SBAC) algorithm that works well for data with mixed numeric and nominal features. A similarity measure proposed by D.W. Goodall (1966) for biological taxonomy, that gives greater weight to uncommon feature value matches in similarity computations and makes no assumptions about the underlying distributions of the feature values, is adopted to define the similarity measure between pairs of objects. An agglomerative algorithm is employed to construct a dendrogram, and a simple distinctness heuristic is used to extract a partition of the data. The performance of the SBAC algorithm has been studied on real and artificially-generated data sets. The results demonstrate the effectiveness of this algorithm in unsupervised discovery tasks. Comparisons with other clustering schemes illustrate the superior performance of this approach.

This paper presents a Bayesian clustering methodology that partitions temporal data into homogeneous groups, and constructs state based profiles for each group in the hidden Markov model (HMM) framework. We propose a Bayesian HMM clustering methodology that improves upon existing HMM clustering algorithm by incorporating HMM model size selection into the clustering control structure. Experimental results indicate the effectiveness of our methodology.

Model based approaches to diagnosis for dynamic systems have been based on continuous and discrete event models. Systems that combine continuous and discrete behaviors, i.e., hybrid systems have been typically abstracted into discrete event models or approximated by continuous models with steep slopes so that existing algorithms can be applied for fault isolation tasks. This approach runs into problems when both discrete events and continuous behaviors provide vital diagnostic information. We propose a diagnostic methodology that uses hybrid models of the system to perform diagnosis.

As robotics and automation applications extend to the service sector, researchers have to increasingly deal with performing robotic actions in uncertain and unstructured environments. A traditional solution to this problem models uncertainty about the effects of actions by probabilities conditioned on the state of the environment, making it possible to select plans that have the highest probability of success in a given situation. Reactive systems use another approach to handling uncertainty, by employing a set of predefined situation-response rules that make it possible to move toward the goal from any situation, whether expected or unexpected. This paper describes a planner that combines the two approaches. A proactive component generates plans that are biased toward picking the most reliable action in a given situation, and a reactive component can alter the selected actions based on unexpected situations that may arise in uncertain environments. Action selection is driven by a spreading activation mechanism on a probabilistic network that encodes the domain knowledge. A decision-theoretic framework incorporates quantitative goal utilities and action costs into the action selection mechanism. Experiments conducted demonstrate the ability of the planner to plan with hard and soft domain constraints and action costs, modify plans as a reaction to unexpected changes in the environment or goal utilities, and plan in situations with multiple conflicting goals.

The complexity of present day embedded systems (continuous processes controlled by digital processors), and the increased demands on their reliability motivate the need for monitoring and fault isolation capabilities in the embedded processors. This paper develops monitoring, prediction, and fault isolation methods for abrupt faults in complex dynamic systems. The transient behavior in response to these faults is analyzed in a qualitative framework using parsimonious topological system models. Predicted transient effects of hypothesized faults are captured in the form of signatures that specify future faulty behavior as higher order time-derivatives. The dynamic effects of faults are analyzed by a progressive monitoring scheme till transient analysis mechanisms have to be suspended in favor of steady state analysis. This methodology has been successfully applied to monitoring of the secondary sodium cooling loop of a fast breeder reactor.

The data exploration task can be divided into three interrelated subtasks: 1) feature selection, 2) discovery, and 3) interpretation. This paper describes an unsupervised discovery method with biases geared toward partitioning objects into clusters that improve interpretability. The algorithm ITERATE employs: 1) a data ordering scheme and 2) an iterative redistribution operator to produce maximally cohesive and distinct clusters. Cohesion or intraclass similarity is measured in terms of the match between individual objects and their assigned cluster prototype. Distinctness or interclass dissimilarity is measured by an average of the variance of the distribution match between clusters. The authors demonstrate that interpretability, from a problem-solving viewpoint, is addressed by the intraclass and interclass measures. Empirical results demonstrate the properties of the discovery algorithm and its applications to problem solving.

This paper discusses systematic methods for diagnosis of complex engineering systems combining qualitative and quantitative analysis of analytic constraint equation system models to generate more precise and accurate candidates. Candidates generated from a qualitative steady-state partial explanation model are refined with available ordinal information on the magnitude of component parameter deviations. In addition, an incremental algorithm is implemented to efficiently process sequences of measurements. Empirical analysis demonstrates that accuracy and resolution of minimal candidate generation are improved by including ordinal information. This avoids the practical problems encountered when reasoning with pure quantitative information.

This paper discusses the design and implementation of PLAYMAKER, a knowledge-based system for characterizing hydrocarbon plays. PLAYMAKER is a component of XX (eXpert eXplorer), a workstation-based tool that aids exploration geologists in a number of different tasks: sediment and carbonate simulation, play and field characterization, retrieval and storage of information in a geological database, comparison of the play or field under study with other fields in the database, and report generation. PLAYMAKER is implemented using MIDST (Mixed Inferencing Dempster-Shafer Tool), a rule-based expert system shell that incorporates mixed-initiative and inexact reasoning based on the Dempster-Shafer evidence combination scheme. This paper discusses the effectiveness of a two-level knowledge base structure adopted for the design and implementation of PLAYMAKER.

Risk assessment of technological processes (chemical and power plants, electro-mechanical systems) is a complex process that requires enumeration of all possible failure modes, their probability of occurrence, and their consequences. Traditionally such studies have been performed by a committee of expert engineers with diverse backgrounds. This paper discusses the use of qualitative modeling techniques based on deriving behavior from structural descriptions and causal reasoning to aid automating and enhancing the risk analysis process. Hierarchical schemes are used for describing component structure, and system functionality is derived from a set of primitive functions and parameters defined for the domain. The system uses these models to automatically generate fault and event networks for hypothesized fault situations specified by users.

No abstract available.

A multi-level system which utilizes both an evidential and a qualitative model for diagnosing fault symptoms in complex electro-mechanical systems is presented. The operation of both models, enhancement by a historical database, and the global control strategy are all discussed. In addition, the constraining of qualitative reasoning with information from the evidential session and the enhancement of the evidential model with information from the qualitative model is demonstrated.

No abstract available.

A framework for qualitative reasoning is discussed that can be applied to a wide variety of engineering tasks such as analysis, design, diagnosis, monitoring, and instruction. A combined component-based and process-theory approach is adopted as the basis for system modeling. This provides a suitable framework for deriving and explaining system behavior in terms of individual component functionality. The current prototype is applied to a space-station waste management system. Examples that deal with the vacuum pump component demonstrate that the methodology can be used to derive both normal and deviant behavior of the system.>

Traditional reasoning methods have found limited success in the medical domain due to their brittleness and lack of robust justification. Recent work has turned to model-based reasoning techniques to overcome these limitations. These techniques employ deep domain models and focus on qualitative reasoning. A discussion is presented of an ongoing project to develop a cardiovascular model with sophisticated reasoning mechanisms and robust explanation capabilities. The framework for the model combines J. DeKleer and J.S. Brown's (1984) component-oriented methods and K.D. Forbus's (1984) process-oriented ontologies. The current prototype models the electrical subsystem of the heart, and simulates normal and deviant cardiac rhythms.>

In this article we describe the conceptual model and processing of (constrained) natural-language queries in information retrieval systems. A language interface based on fuzzy set techniques is proposed to handle the uncertainty inherent in natural-language semantics. The conceptual model is developed and exemplified in the context of document retrieval. Specifically, the user query is considered to be a triple, q = (qc, qy, qn), where qc indicates the part of the query that deals with concepts and operators that link these concepts, qy identifies the publication period the user is interested in, and qn, pertains to the number of documents to be retrieved. We describe query decomposition using an augmented transition network parser and the assignment of functions and relations needed by each portion of the query to represent uncertainties inherent in the natural language. The output of the natural-language interface is then passed to a knowledge-based retrieval mechanism that will be described in a companion article (Part II). © 1987 John Wiley & Sons, Inc.

This article presents our conceptual model of the retrieval process of a document-retrieval system. The retrieval mechanism input is an unambiguous intermediate form of a user query generated by the language processor using the method described previously. Our retrieval mechanism uses a two-step procedure. In the first step a list of documents pertinent to the query are obtained from the document database, and then an evidence-combination scheme is used to compute the degree of support between the query and individual documents. The second step uses a ranking procedure to obtain a final degree of support for each document chosen, as a function of individual degrees of support associated with one or more parts of the query. The end result is a set of document citations presented to the user in ranked order in response to the information request. Numerical examples are given to illustrate various facets of the overall system, which has been prototypically implemented in modular form to test system response to changes in model parameters. © 1987 John Wiley & Sons, Inc.

An application of knowledge-based systems in the field of business decisionmaking is presented. The expert's reasoning processes and problem-solving techniques in the operations analysis domain are the primary focus, and OASES, an operations analysis expert system, is designed to play the role of an intelligent assistant and to aid management in diagnosing problems in production processes, such as automobile assembly or textile manufacturing. The system combines forward and backward inferencing mechanisms to interact with users in a mixed initiative format. The knowledge base is designed as a partitioned rule base, and the inference engine uses the Dempster-Shafer scheme for inexact reasoning.

A knowledge-based system approach is applied to the design of a document retrieval system for online retrieval of bibliographic material. The main components of the system are a language interface that understands user queries in pseudo natural language (i.e. queries that are restricted to subject domain concepts), a retrieval component that combines the expertise of subject-domain and retrieval experts, a help system that aids users in formulating and reformulating queries, and a user model builder that infers specific user characteristics. A comprehensive mathematical model based on fuzzy sets and fuzzy relations has been defined for the thesaurus and the retrieval mechanism and a prototypical system has been implemented in a modular fashion to test system response to changes in model parameters.

A number of linear and nonlinear mapping algorithms for the projection of patterns from a high-dimensional space to two dimensions are available. These two-dimensional representations allow quick visual observation of a data set. A combination of two popular mapping algorithms-Sammon's mean-square error technique and the triangulation method-is proposed to overcome the limitations in the individual algorithms. Some factors which describe the goodness of a projection are described, and a comparison is made of six of these algorithms by running them on four data sets. The results obtained support the use of the proposed algorithm.