Model-Based Processes for Real-Time Cyber-Physical Systems Testbeds

Jonathan Sprinkle
University of Arizona
Electrical and Computer Engineering

This talk discusses sensing models and realistic control techniques that can be used within CPS testbeds across simulation, software-in-the-loop, and hardware-in-the-loop portions of a research project’s design cycle. The talk will include analysis and design of controllers for the physical system testbed (a full-sized Ford Escape). The discussion will introduce the open-source simulation tools, and demonstrate how the model-based approach taken in control and sensing permits an agile design cycle—in contrast to a traditional waterfall cycle. For case study examples, details of two high-profile experiments using the CAT Vehicle will be provided, and the process by which the Robotics System Toolbox is integrated into the experiments that (i) showed an ability to improve traffic flow using a single autonomous vehicle in the flow, and (ii) exercised adaptive cruise control models to determine potential instabilities.