Model-Based Design Approach for Validation of Vehicle Longitudinal Control Algorithm

Abstract

In this paper, model-based testing strategies are described for the validation of an Adaptive Cruise Control (ACC) algorithm developed for a 2019 Chevrolet Blazer as part of the EcoCAR Mobility Challenge. A team of undergraduate and graduate students developed testing procedures to assess model fidelity, and to identify and resolve issues with the algorithm before deployment to a student-modified production vehicle. The algorithm validation is conducted via three progressive levels of validation environments: Model-In-the Loop (MIL), Hardware-In-the-Loop (HIL), and Driver-In-the-Loop (DIL). When the ACC algorithm is evaluated using system requirements in the testing sequence, the MIL environment performs the tests at least 87% faster than the HIL environment. The MIL environment can also utilize parallel computing, which leverages multi-core CPUs to conduct multiple simulations simultaneously. Although comparisons between MIL and HIL results revealed good agreements, slight differences in system dynamics highlights a need for future Vehicle-In-the-Loop (VIL) testing. By showing how the concepts can be applied to the validation of an autonomous feature in a vehicle with detailed test scenarios and evaluation metrics, the paper will serve as a good reference for the students and engineers interested in this field.