Simulation-Based Testing and Evaluation of Automated Vehicle Motion Planning and Control Algorithms
Özet
Automated vehicles can help to reduce accidents, stress on the driver and the ecological footprint of vehicles by utilizing motion planning and control algorithms. Ensuring the safety and comfort of drivers and passengers in automated vehicles requires delicate testing and verification process of newly developed algorithms. This study aims to establish an exchangeable and adjustable MATLAB® simulation environment with a flexible architecture to develop motion planning and control algorithms; and then test and evaluate the algorithms in conditions such as flowing traffic and different tire-road coefficient of frictions in compliance with the Automated Lane Keeping System regulations; to detect the design imperfections and identify the critical scenarios. The study yielded a Motion Planning Algorithm, a Simulation Environment in MATLAB® to test and evaluate algorithms, and an Automated Scenario generation and execution to identify critical scenarios and design flaws. Algorithms and parameters used in this research can be modified to perform sensitivity analysis for individual effects and contributions.