Comparison of Model-Predictive control and PID control for Adaptive Cruise Control of UW EcoCAR vehicle

Abstract

This work evaluates two control strategies for Adaptive Cruise Control (ACC), classical control (PID control), and Model Predictive Control (MPC) with linear-piecewise approximated engine fuel map as a part of cost function to penalize fuel consumption, both applied to UW EcoCAR vehicle model. The ACC system with MPC consists of hierarchical control architecture, a lower-level controller to track the acceleration command, and a higher-level ACC control. The control algorithms are tested on Model-In-Loop (MIL) using Simulink. The real-time Hardware-In-Loop (HIL) performance testing is done using the dSpace simulator which runs the vehicle model and the dSpace MicroAutoBox II which serves as a controller platform. A comparison of miles per gallon of fuel, average acceleration, and average jerk is provided for drive cycle runs by ACC PID and MPC.