This project explores new mathematical techniques that provide a scientific basis to understand the fundamental properties of Cyber-Physical Systems (CPS) controlled by Artificial Intelligence (AI) and guide their design. From simple logical constructs to complex deep neural network models, AI agents are increasingly controlling physical/mechanical systems. Self-driving cars, drones, and smart cities are just examples of AI-controlled CPS. However, regardless of the explosion in the use of AI within a multitude of CPS domains, the safety and reliability of these AI-controlled CPS is still an under-studied problem. This project includes activities integrated with education, so as to explore how learning through counterexamples works for AI, and to help with critical thinking skills for young students.

The goal of this project is the advancement of machine learning dynamic models and real-time control systems for human cyber-physical balance systems. Ranging from biped walkers and human bicycle riding to human-controlled helicopters, human cyber-physical balance systems maintain challenging tasks of simultaneously trajectory-tracking and unstable platforms balancing. Although many physical models were developed in past decades, it is still challenging to safely and effectively operate these human-in-the-loop balance machines in highly variable, uncertain environments. This project will develop machine learning-based mathematical models and robust control strategies for human cyber-physical balance systems. The researchers will also develop a number of integrated research and education programs to attract students from underrepresented groups into engineering and involve undergraduate students into research. Human cyber-physical balance systems involve human movements as physical and forceful interactions with unstable, underactuated platforms. It is challenging to capture and control physical human-machine or human-robot interactions in complex, uncertain environments. This project will focus on: (1) development of machine learning-based models and characterization for human cyber-physical balance systems; (2) development of new hardware/software co-design accelerated learning-based real-time control to handle human cyber-physical balance system dynamics in highly variable, uncertain environments; and (3) robotic testbeds development, experimental validation and performance evaluation. The integration of data-driven model and learning-based control strategies, along with the hardware/software co-design enabled real-time implementation, provides new perspectives on performance enhancement of safety-critical or mission-critical cyber-physical systems in dynamic, uncertain environments.

Artificial Intelligence (AI) has shown superior performance in enhancing driving safety in advanced driver-assistance systems (ADAS). State-of-the-art deep neural networks (DNNs) achieve high accuracy at the expense of increased model complexity, which raises the computation burden of onboard processing units of vehicles for ADAS inference tasks. The primary goal of this project is to develop innovative collaborative AI inference strategies with the emerging edge computing paradigm. The strategies can adaptively adjust cooperative inference techniques for best utilizing available computation and communication resources and ultimately enable high-accuracy and real-time inference. The project will inspire greater collaborations between experts in wireless communication, edge computing, computer vision, autonomous driving testbed development, and automotive manufacturing, and facilitate AI applications in a variety of IoT systems. The educational testbed developed from this project can be integrated into courses to provide hands-on experiences. This project will benefit undergraduate, master, and Ph.D. programs and increase under-represented groups? engagement by leveraging the existing diversity-related outreach efforts. A multi-disciplinary team with complementary expertise from Rowan University, Temple University, Stony Brook University, and Kettering University is assembled to pursue a coordinated study of collaborative AI inference. The PIs explore integrative research to enable deep learning technologies in resource-constrained ADAS for high-accuracy and real-time inference. Theory-wise, the PIs plan to take advantage of the observation that DNNs can be decomposed into a set of fine-grained components to allow distributed AI inference on both the vehicle and edge server sides for inference acceleration. Application-wise, the PIs plan to design novel DNN models which are optimized for the cooperative AI inference paradigm. Testbed-wise, a vehicle edge computing platform with V2X communication and edge computing capability will be developed at Kettering University GM Mobility Research Center. The cooperative AI inference system will be implemented, and the research findings will be validated on realistic vehicular edge computing environments thoroughly. The data, software, and educational testbeds developed from this project will be widely disseminated. Domain experts in autonomous driving testbed development, intelligent transportation systems, and automotive manufacturing will be engaged in project-related issues to ensure relevant challenges in this project are impactful for real-world applications.

This Faculty Early Career Development Program (CAREER) award will contribute to the advancement of national prosperity and economic welfare by researching systems that improve access to manufacturing services. Wearable electronics are widely used in health monitoring and wearable computing and there is a compelling need for comfort, biocompatibility, and easy operation. Recent progress in smart fabrics, textiles, and garments and the associated manufacturing technologies provides opportunities for next-generation wearable electronic devices that are fabricated on cloth. Automatic embroidery manufacturing is now an accessible tool for individuals and entrepreneurs. Embroidery offers great potential for electronic design due to its flexibility in transferring a desired pattern to fabric substrates. This project aims to establish a cloud manufacturing framework that integrates electronics and design-to-manufacturing translation in a system that can be used by customers, manufacturers, design experts, and developers to design and produce embroidered wearable electronics. In addition, this project also aims to broaden participation from K-12, undergraduate, and graduate students, to provide rich multidisciplinary classroom and non-classroom experiences for all levels of students, and to inspire student interest in STEM careers.

For mission-critical Cyber-physical systems (CPSs), it is crucial to ensure these systems behave correctly while interacting with open, dynamic, and uncertain environments. Synthesizing CPSs with assurance is a daunting task: On the one hand, the interconnected networks, sensors, and (semi-) autonomous systems introduce unprecedented vulnerabilities to both cyber- and physical spaces; On the other hand, purposeful attacks may aim to compromise more complex system properties beyond traditional stability and safety. For example, in a robotized security patrol system, a successful cyber-attack on the sensor network can be combined with adversarial control commands to compromise the system and disrupt its mission. In this project, the goal is to develop intelligent sensing and control methods for CPSs that leverage advanced cyber defense techniques for constructing provably secured systems subject to high-level complex mission objectives. To achieve this goal, this project will develop formal modeling and solutions for a class of CPS Games, featured by multi-stage, strategic interactions between a controller/defender and a (coordinated cyber- and physical-) attacker. The synthesis of CPS control/defense strategies will explicitly account for asymmetric information and investigate how to leverage cyber defense and deception for guaranteed performance in mission-critical CPSs. The fundamental theory and algorithms will be validated via both a physical testbed including multiple mobile robots and a wireless sensor network and a simulation as a virtual proving ground. The technical contributions are: 1) to develop solutions of imperfect information games with temporal logic objectives and employ the solutions to design control and information acquisition strategies for CPSs under attacks; 2) to investigate the use of cyber defense techniques for gaining defender information advantages against CPS attackers in the synthesis of integrated cyber-physical defense; 3) to develop control strategies against coordinated cyber- and physical- attacks, and novel control methods that incorporate human?s preference for CPS resilience against prior unknown attacks. The project also includes an educational plan by developing an interdisciplinary curriculum for engineering students on cyber-physical security and attack mitigation techniques. The PI will promote K-12 outreach by participating in campus-wide activities and offering a two-week summer course for high-school students with hands-on robotic and wireless sensor networks experiments, aiming to raise students? interest in science and engineering and public?s awareness of CPS security.

The goals of Automated Driving Systems (ADS) and Advanced Driver Assistance Systems (ADAS) include reduction in accidental deaths, enhanced mobility for differently abled people, and an overall improvement in the quality of life for the general public. Such systems typically operate in open and highly uncertain environments for which robust perception systems are essential. However, despite the tremendous theoretical and experimental progress in computer vision, machine learning, and sensor fusion, the form and conditions under which guarantees should be provided for perception components is still unclear. The state-of-the-art is to perform scenario-based evaluation of data against ground truth values, but this has only limited impact. The lack of formal metrics to analyze the quality of perception systems has already led to several catastrophic incidents and a plateau in ADS/ADAS development. This project develops formal languages for specifying and evaluating the quality and robustness of perception sub-systems within ADS and ADAS applications. To enable broader dissemination of this technology, the project develops graduate and undergraduate curricula to train engineers in the use of such methods, and new educational modules to explain the challenges in developing safe and robust ADS for outreach and public engagement activities. To broaden participation in computing, the investigators target the inclusion of undergraduate women in research and development phases through summer internships.

Active user participation in large-scale infrastructure systems, while presenting unprecedented opportunities, also poses significant challenges for the operator. One such example is electric power distribution systems, where the massive integration of distributed energy resources (DERs) and flexible loads motivates new decision-making paradigms via demand response through user engagement. This project introduces a novel approach for intelligent decision making in power distribution systems to efficiently leverage flexible demand commitments in highly uncertain and stochastic environments. The project goals are to (1) develop analytics required to enable actionable demand-side flexibility from several small consumers by adequately representing their constraints regarding electricity usage and their interactions with the system and the energy provider; and (2) develop a prototype for demand-side coordination using an open-source testbed for distribution systems management and evaluate the proposed algorithms with real-world utility data. Successful completion of this project will provide solutions to adaptive and smart infrastructure systems in which passive users turn into active participants. For the demand response focus here, this project will enable high levels of penetration of flexible loads and DERs economically through the transformation of grid operation from load following to supply following. The results from this project will provide valuable guidance to policymakers and electric utilities in managing aggregator-driven markets. The central aim of this proposal is to enable the demand-side participation of many small customers in a distribution grid and solve for an interface between customers and an energy provider. The proposed architecture follows a two-level structure: a home energy management system (HEMS) providing a home-level interaction between the consumer and the HEMS, and a feeder-level interaction between the HEMS and the demand-response provider. Research along two thrusts will be proposed: (1) learning-based control to achieve home-level flexibility upon learning and incorporating customer constraints and preferences into the decision-making process; and (2) game-theoretic constructs to aggregate and coordinate the home-level flexibility at the network-level in a constrained environment with unknown customer utility functions. Technical innovations at the HEMS-customer interface will include automata learning-based algorithms used by HEMS to learn customers? temporally evolving energy usage constraints, and reinforcement learning algorithms to satisfy temporal constraints while optimizing the cost of electricity consumption. At the provider-HEMS interface, technical innovations will include a new mean field based model of customers that allows the provider to interact with only a few customer classes, and a Stackelberg game formulation that explicitly incorporates network congestion constraints.

Shared electric micromobility (SEM) services such as shared electric bikes and scooters, as an emerging example of mobile cyber-physical systems, have been increasingly popular in recent years for short-distance trips such as from bus stops to home, enabling convenient mobility through multi-modal transportation and less environmental impact by reducing emission by traffic congestion. However, the success of the service depends on the effective and efficient management of thousands of electric vehicles (e.g., bikes or scooters). Existing management frameworks mainly focused on balancing demand and supply considering energy recharging. However, most of them, if not all, ignored human interactions with systems (e.g., how users select and use vehicles), which leads to a significant gap between experimental and real-world effectiveness. The objective of this project is to develop an interaction-aware management framework for mobile cyber-physical systems.

The use of artificial intelligence in cyber-physical systems is limited by challenges such as data availability, task environment complexity, and the need for expressive and interpretable high-level knowledge representations. To address these challenges, this project aims to develop a set of neuro-symbolic learning and control tools by integrating machine learning, control theory, and formal methods. The results are expected to find application across cyber-physical systems such as robotic systems, autonomous systems, and networked cyber-physical systems. Validation in a testbed environments should facilitate safe deployments in real-world physical environments with provable guarantees and robustness against potential adversaries.

The gradual deployment of self-driving cars will inevitably lead to the emergence of a new important class of cyber-physical-human systems where autonomous vehicles interact with human-driven vehicles via on-board sensors or vehicle-to-vehicle communications. Reinforcement learning along with control theory can help meet the safety requirements for real-time decision making and Level 5 autonomy in self-driving vehicles. However, it is widely known that conventional reinforcement learning policies are vulnerable to adversarial or non-adversarial perturbations to their observations, similar to adversarial examples for classifiers and/or reward (packet) drops of the learning. Such issues are exacerbated by concerns of addressing resiliency as the use of open communication and control platforms for autonomy becomes essential, and as the industry continues to invest in such systems. Decision making mechanisms, designed to incorporate agility with the help of reinforcement learning, allow self-adaptation, self-healing, and self-optimization. This research will contribute and unify the body of knowledge of several diverse fields including reinforcement learning, security, automatic control, and transportation for resilient autonomy with humans-in-the-loop.