Design and Formal Verification of an Adaptive Cruise Control Plus (ACC+) System

Abstract

Stop-and-Go Adaptive Cruise Control (ACC+) is an extension of Adaptive Cruise Control (ACC) that works at low speed as well as normal highway speeds to regulate the speed of the vehicle relative to the vehicle it is following. In this thesis, we design an ACC+ controller for a scale model electric vehicle that ensures the robust performance of the system under various models of uncertainty. We capture the operation of the hybrid system via a state-chart model that performs mode switching between different digital controllers with additional decision logic to guarantee the collision freedom of the system under normal operation. We apply different controller design methods such as Linear Quadratic Regulator (LQR) and H-infinity and perform multiple simulation runs in MATLAB/Simulink to validate the performance of the proposed designs. We compare the practicality of our design with existing formally verified ACC designs from the literature. The comparisons show that the other formally verified designs exhibit unacceptable behaviour in the form of mode thrashing that produces excessive acceleration and deceleration of the vehicle. While simulations provide some assurance of safe operation of the system design, they do not guarantee system safety under all possible cases. To increase confidence in the system, we use Differential Dynamic Logic (dL) to formally state environmental assumptions and prove safety goals, including collision freedom. The verification is done in two stages. First, we identify the invariant required to ensure the safe operation of the system and we formally verify that the invariant preserves the safety property of any system with similar dynamics. This procedure provides a high level abstraction of a class of safe solutions for ACC+ system designs. Second, we show that our ACC+ system design is a refinement of the abstract model. The safety of the closed loop ACC+ system is proven by verifying bounds on the system variables using the KeYmaera verification tool for hybrid systems. The thesis demonstrates how practical ACC+ controller designs optimized for fuel economy, passenger comfort, etc., can be verified by showing that they are a refinement of the abstract high level design.