Trustworthy Autonomy

3 units

Cyber-physical systems are becoming increasingly capable. These are systems consisting of digital and physical components. They are deployed to increase autonomy in critical settings where failure is costly, such as driving, aviation, medicine, and manufacturing. This course covers several approaches for assuring that such systems meet their specifications. We will discuss: (1) how to mathematically model these systems and their specifications, (2) how to formally verify and synthesize models which meet their specifications, and (3) how to address the challenges resulting from incorporating machine-learned components, such as neural networks, as perception and control components in these systems.