Cyber-physical systems consist of computers that sense and control a rapidly changing physical environment. Examples of such systems include autonomous transportation (self-driving cars), traffic networks, energy distribution, power networks, air traffic control and integrated medical devices. Recently, advances in artificial intelligence and robotics have given rise to autonomous (cyber-physical) systems that make decisions in response to environment changes, often without human intervention. Cyber-Physical Systems are often safety-critical. Failure of these systems can result in loss of human lives. Therefore, the key challenge is to engineer them so that people can trust them with their lives. As a result, providing rigorous guarantees about the correctness and performance of these systems is of extreme importance.
Our group works on fundamental problems that integrate ideas from theory of computation, algorithms, programming languages, dynamics and control, convex optimization, game theory and embedded systems.
We also collaborate on problems of deep societal interest with researchers in AI/robotics (automatically synthesizing controllers for self-driving cars), human-computer interaction (helping people learn fly UAVs without crashing them), aerospace engineering (learning requirements for autonomous UAV missions) and the UC Anschutz medical school (treating type-1 diabetes using autonomous medical devices).
Current topics of interest include:
- Formal verification and control of hybrid dynamical systems
- Runtime monitoring for autonomous systems
- Safety of AI-enabled cyber-physical systems
- Safety critical medical devices: implantable defibrillators and artificial pancreas.
- Distributed optimization and control
- Game theory
- Reinforcement learning
- Learning and sequential decision-making
- CSCI 5854: Foundations of Autonomous Systems
- CSCI 5654: Linear and Integer Programming
- CSCI 7135: Probabilistic Programming
- CSCI 5254: Convex Optimization and Its Applications