Intersection of Machine Learning with Control Control Theory Fundamentals of Multi-Agent Systems Safe Motion Planning and Control for Autonomous Driving Under Multi-Source Uncertainty Published Special Sections Intersection of Machine Learning with Control (Recurring) Editor: Kyriakos G. Vamvoudakis Georgia Institute of Technology https://kyriakos.ae.gatech.edu [email protected] Submission Window: 15 Oct 2024 – 15 April 2025 Extended through 31 May 2025! The section will be recurrent, and a planned submission window will be open at least once a year. Harnessing the power of machine learning to continuously monitor and detect anomalies advances the state of the art in instrumentation control. Learning-enabled systems have been rapidly increasing in size and acquiring new capabilities. These systems are typically deployed in complex operating environments, so their safety becomes extremely important. Ensuring safety requires that systems are robust to extreme events while we can monitor them for anomalous and unsafe behavior. While traditional machine learning systems are evaluated pointwise with respect to a fixed test set, such static coverage provides only limited assurance when exposed to unprecedented conditions in complex operating environments. One key question that remains unanswered is “How can we design and deploy learning-enabled systems that can be robust to extreme events while monitoring them for anomalous and unsafe behavior?” This special issue aims to contribute to this growing area of interest and thus calls for papers in this topical area. Topics of interest for this special issue include and are not limited to: Machine learning for dimensionality reduction and system identification Emerging applications for learning-based control Data-driven optimization and control for dynamical systems Safe reinforcement learning and safe adaptive control Bridging model-based and learning-based control systems Distributed learning over distributed systems Reinforcement learning for multiagent systems Optimization, dynamics and control for machine learning Reinforcement learning and statistical learning for dynamical and control systems Please contact the OJ-CSYS editorial assistant with any questions. Special Section on Intersection of Machine Learning with Control Download Control Theory Fundamentals of Multi-Agent Systems Editors: David Casbeer Air Force Research Laboratory [email protected] Dejan Milutinović University of California, Santa Cruz [email protected] Tansel Yucelen University of South Florida [email protected] Submission Window: 15 Feb 2025 – 15 April 2025 Extended through 31 May 2025! As technology develops to allow integration of mobility, computing, and communication on single platforms, we have entered a new era in which teams of agents, physical robots, or sets of control laws interact with each other to influence their states, motions or actions to cooperatively perform tasks in an array of civilian and military applications. The topics of this special section are control theory and experimental results addressing the emergence of desirable outcomes in multi-agent, robotic, or complex control systems, as well as fundamental limitations for achieving the emergence. Results based on nonlinear control, optimal control, game theory, and other areas of control theory and machine learning are also in the scope of the section. Prospective authors are invited to submit original contributions addressing two or more agents with a clear relevance to multi-agent systems. Topics of interest for this special issue include and are not limited to: Distributed control under spatiotemporal constraints Coordination and motion planning for multi-agent systems Coordination and collaboration in heterogeneous multi-agent systems Cooperative manipulation Game theory approaches to multi-vehicle systems Resilient decision-making for autonomous multi-agent systems Machine learning approaches for multi-agent systems Control or machine learning approaches for control action synergies Distributed decision-making Optimization in multi-agent systems Please contact the OJ-CSYS editorial assistant with any questions. Special Section on Control Theory Fundamentals of Multi-Agent Systems Download Safe Motion Planning and Control for Autonomous Driving Under Multi-Source Uncertainty Editor: Peter Seiler University of Michigan [email protected] https://seiler.engin.umich.edu Organizers: Zejiang Wang University of Texas at Dallas [email protected] Yuanqiu Mo Southeast University [email protected] Mingyu Cai University of California, Riverside [email protected] Jinhao Liang National University of Singapore [email protected] Guodong Yin Southeast University [email protected] Submission Window: 28 Feb 2025 – 31 August 2025 With the widespread attention in both academia and industry regarding autonomous driving, ensuring safety in its autonomous planning and control has become an urgent task. Currently, the implementation framework of autonomous driving primarily involves the planner generating a “collision-free trajectory,” which is then handed over to the control layer for tracking. However, real-world predictions and vehicle dynamics are typically fraught with various uncertainties (i.e., the intentions and trajectories of other vehicles, vehicle steering system hysteresis, etc.), especially in scenarios involving high-speed vehicle maneuvering behavior. These uncertainties can undermine safety planning validated by nominal model verification, rendering the control trajectory no longer safe. While it seems that the vast majority of machine learning methods struggle to provide safety assurances for planning and control, some trustworthy ideas for machine learning-based safety planning and control have begun to emerge. Authors are encouraged to focus on the uncertainties encountered in real-world scenarios, with a focus on developing algorithms capable of explicitly reasoning uncertainty propagation and achieving autonomous driving planning and control with strong robustness. This special section advocates for researchers to attentively address the multifaceted challenges inherent in the domain of autonomous driving planning and control in order to provide apt solutions. Topics of interest for this special issue include and are not limited to: Robust trajectory planning of autonomous driving Robust trajectory tracking control of autonomous driving Safety-critical critical decision-making and control of autonomous driving Robust and resilient autonomous vehicle control subject to sensor failure or attack Machine learning for trajectory prediction Statistical methods for quantifying uncertain predictions Human-centered AI technology for autonomous driving control Platoon control subject to various heterogeneity Application of game theory in the human-vehicle interaction system Cooperative control between human-driven (or human-operated) vehicles and autonomous vehicles Please contact the OJ-CSYS editorial assistant with any questions. Special Section on Safe Motion Planning and Control for Autonomous Driving Under Multi-Source Uncertainty Download Published Special Sections OJ-CSYS Vol. 3 Resilient and Safe Control in Multi-Agent Systems Multi-agent systems (MASs) have gained a lot of popularity in recent years in different disciplines as a means to solve complex tasks by subdividing them into smaller problems. The modular and interconnected structure provides important benefits, like scalability and design flexibility, but can also lead to vulnerability, by allowing local faults to spread to neighboring locations and even to the whole system. This challenge is outside the scope of robust and adaptive control, where the controller typically assumes prior knowledge about disturbances affecting the model. This is especially true for adversarial disturbances, which cannot be modeled with confidence and may fatally disrupt a control task at the global level. Nonetheless, propagating failures are a crucial issue in several application domains, from power grids subject to cyber-attacks and power outages, to multi-robot systems dealing with unexpected changes in the surrounding environment, to vehicular networks experimenting with unpredictable behaviors of human drivers. These cases require resilient strategies that can restore the system functionalities on-the-fly in the face of unexpected or adversarial conditions that fall outside of the design assumptions. A unifying framework for resilience in MASs is still lacking, and the complexity of large-scale systems may prevent applicability of several proposed approaches. Ultimately, the path toward resilient networked control systems is still long, and further research and technological effort is needed to cope with adversities of any kind and increasing sophistication. This special issue aims to contribute to this growing area of interest. Guest Senior Editors Luca Schenato, University of Padova Ruggero Carli, University of Padova Guest Associate Editors Mauro Franceschelli, University of Cagliari Hideaki Ishii, Tokyo Institute of Technology Maryam Kamgarpour, Swiss Federal Institute of Technology Lausanne Yancy Diaz-Mercado, University of Maryland Chuchu Fan, MIT Solmaz Kia, University of California, Irvine Stephen Smith, University of Waterloo Pavankumar Tallapragada, Indian Institute of Science Minghui Zhu, Pennsylvania State University OJ-CSYS Vol. 3 Control and Monitoring of Next-Gen Urban Infrastructure: Applications to Power, Transportation, and Water Systems Climate change, overpopulation, aging infrastructure, urbanization, and the natural finiteness of earth’s resources has pushed urban designers, city planners, policy makers, scientists, and engineers to rethink traditional control and design paradigms, and to look for holistic solutions to ensure the safety, resilience, security, and efficiency of operating new infrastructure. In particular, the three key infrastructure – electric power systems, water systems, and traffic networks – all face monumental challenges related to real-time operation. To that end, this special section focuses on presenting and sharing new control algorithms and architectures for the next generation of urban infrastructure with a specific focus on power, transportation, and water systems. Guest Senior Editor Ahmad Taha, Vanderbilt University Guest Associate Editors Maria Laura Delle Monache, University of California, Berkeley Mads Almassalkhi, University of Vermont Christian Claudel, University of Texas at Austin Ahmed A. Abokifa, University of Illinois Chicago Marcio Giacomoni, University of Texas at San Antonio Mahnoosh Alizadeh, University of California, Santa Barbara Carlos Ocampo-Martinez, Universitat Politecnica de Catalunya OJ-CSYS, Vol. 3 Modeling, Control, and Learning Approaches for Human-Robot Interaction Systems Despite significant advances in robotics and autonomy, human participation is essential for the practical utilization and performance of such systems. An individual may act as a decision-maker, supervisor, or collaborator of a robot, working together to achieve a common goal. The synergy of human intelligence with robots has been shown to improve the joint human-robot system performance and reduce workload. There are many important aspects to enable effective human-robot interaction (HRI), e.g., the design of user-friendly human-machine interfaces and meta-analysis of human factors affecting robot behaviors. Among all these aspects, there is a great demand to perform system-level analysis, estimation, and prediction and provide performance guarantees for these HRI systems. The problem is challenging due to the uncertain nature of human behaviors and interactions with robots. This, therefore, calls for innovations in the modeling, control, and learning approaches and their integrations for HRI systems. Guest Senior Editor Yue Wang, Clemson University Guest Associate Editors Yancy Diaz-Mercado, University of Maryland Victor H. Duenas, Syracuse University Takeshi Hatanaka, Tokyo Institute of Technology Sandra Hirche, Technische Universität München Neera Jain, Purdue University Meeko Oishi, University of New Mexico Stephen L. Smith, University of Waterloo Vaibhav Srivastava, Michigan State University Yildiray Yildiz, Bilkent University OJ-CSYS, Vol. 2 Synchronization in Natural and Engineering Systems Synchronized behaviors arise spontaneously and by design in various natural and man-made systems. For instance, distinctive network-wide patterns of synchrony determine the coordinated motion of orbiting particle systems, promote successful mating in populations of fireflies, regulate the active power flow in electrical grids, and enable numerous cognitive functions in the brain. While some systems rely on synchronization of all units to function properly, other systems exhibit a rich repertoire of synchronized behaviors including cluster synchronization, chimera states, explosive synchronization patterns, and even transient, cross-frequency, and phase-amplitude synchronization. These coordinated behaviors can emerge from the properties of the interconnection structure among the units, be the result of the dynamics of the isolated units, rely on the interplay of structure and dynamics, or be driven by exogenous stimuli. Despite being one of the most studied phenomena in science and engineering, the principles underlying general synchronization patterns in complex systems and, importantly, effective methods to regulate different forms of synchronized behaviors, have remained elusive. Guest Senior Editor Fabio Pasqualetti, University of California, Riverside Guest Associate Editors Vaibhav Srivastava, Michigan State University ShiNung Ching, Washington University in St. Louis Erfan Nozari, University of California, Riverside Giovanni Russo, University of Salerno, Italy Adilson E. Motter, Northwestern University Zahra Aminzare, University of Iowa Madalena Chaves, Inria Sophia Antipolis - Mediterranean Corentin Briat, ETH-Zürich Switzerland OJ-CSYS, Vol. 2 Formal Verification and Synthesis of Cyber-Physical Systems CPSs are complex systems resulting from intricate interactions of computational devices with the physical plants. Recent advances in device manufacturing, computation, and storage have made tremendous advances in hardware and systems platforms for CPSs. With this growing trend in computational devices, CPSs are becoming more and more ubiquitous with many safety-critical applications including autonomous transportations, robot-assisted surgery, medical devices such as artificial pancreas, smart manufacturing, smart buildings, etc. Unfortunately, the analysis and design of CPSs nowadays are still based on ad-hoc solutions sought by simply taking the union of the classical techniques in control theory and computer science. This results in error-prone analysis or design, and very high testing and validation costs. Formal-methods based approach to CPS design recommends rigorous requirement specification in every stage of the system development. Formal verification and controller synthesis are two leading approaches to provide correctness guarantees for CPS with respect to such requirements. While formal verification aims at providing a proof of correctness with respect to the given specifications, the goal of the controller synthesis approach is more ambitious: it takes a control system together with the specification and produces a controller such that the resulting closed-loop satisfies the specification. Guest Senior Editor Majid Zamani, University of Colorado, Boulder Guest Associate Editors Samuel Coogan, Georgia Institute of Technology Melkior Ornik, University of Illinois Urbana-Champaign Chuchu Fan, MIT Ebru Aydin Gol, Middle East Technical University Raphael Jungers, UCLouvain Jun Liu, University of Waterloo Meeko Oishi, University of New Mexico Jana Tumova, KTH Royal Institute of Technology Anne-Kathrin Schmuck, Max Planck Institute for Software Systems Abolfazl Lavaei, Newcastle University OJ-CSYS, Vol. 1, 2, 3 Intersection of Machine Learning with Control Unprecedented technological advances have fueled the creation of devices that can collect, generate, store, and transfer large amounts of data. This massive data outpour is profoundly changing the way in which complex engineering problems are solved, calling for the conception of new interdisciplinary tools at the intersection of machine learning, dynamic systems and control, and optimization. While the repurposing of control theories building on new Machine Learning methods can be highly successful, Dynamic Systems and Control can greatly contribute to analyze and devise novel adaptive, safety-critical controllers with performance guarantees. Vol. 3 Guest Senior Editor Peter J. Seiler, University of Michigan Guest Associate Editors Neera Jain, Purdue University Wang Gang, Beijing Institute of Technology Carlos Ocampo-Martinez, Universitat Politècnica de Catalunya Huazhen Fang, University of Kansas Insoon Yang, Seoul National University Alberto Speranzon, Honeywell Aerospace Minghui Zhu, Pennsylvania State University Yue Wang, Clemson University Mahnoosh Alizadeh, University of California, Santa Barbara Vol. 2 Guest Senior Editor Lacra Pavel, University of Toronto Guest Associate Editors Neera Jain, Purdue University Alberto Speranzon, Lockheed Martin Wang Gang, Beijing Institute of Technology Minghui Zhu, Pennsylvania State University Somayeh Sojoudi, University of California, Berkeley Yue Wang, Clemson University Peter Seiler, University of Michigan Mahnoosh Alizadeh, University of California, Santa Barbara Insoon Yang, Seoul National University Vol. 1 Senior Editor Sonia Martinez, University of California, San Diego Guest Associate Editors Neera Jain, Purdue University Massimo Canale, Politecnico di Torino, Turin Somayeh Sojoudi, University of California, Berkeley Peter Seiler, University of Michigan Huazhen Fang, University of Kansas Insoon Yang, Seoul National University Alberto Speranzon, Lockheed Martin Minghui Zhu, Pennsylvania State University Yue Wang, Clemson University Mahnoosh Alizadeh, University of California, Santa Barbara
