Motion planning and control for autonomous vehicle collision avoidance systems using potential field-based parameter scheduling

• Published in: Machine Intelligence in Mechanical Engineering
• Main Author: Wahid N.; Zamzuri H.; Amer N.H.; Dwijotomo A.; Saruchi S.A.
• Format: Book chapter
• Language: English
• Published: Elsevier 2024
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85191131015&doi=10.1016%2fB978-0-443-18644-8.00003-4&partnerID=40&md5=b4f9aa93011d787b3dd8efbd115f2409
• DOI: 10.1016/B978-0-443-18644-8.00003-4

Establishing an efficient and safe maneuver is an important part toward the successful development of autonomous vehicle collision avoidance systems in encountering the risk of imminent collision. A real driving environment deals with various dynamic conditions such as different vehicle speeds and numerous driving situations. Therefore, an adaptive strategy in a collision avoidance system is necessary in providing an appropriate vehicle motion and feasible trajectory of control for collision-free maneuver to guarantee safety. This study proposed a motion planning and control strategy for an autonomous vehicle collision avoidance system based on the potential field (PF) approach with a combination of the parameter scheduling technique. A particle swarm optimization algorithm is used to optimize the knowledge database information that is developed based on the perception of driver toward risk in the driving environment. This is the main component in developing the adaptive mechanism to adapt to numerous vehicle speeds and different obstacle positions during avoidance maneuver. The main contribution of this work is the improvement of a feasible vehicle motion for safe collision avoidance maneuver that is generated based on the reference lateral motion provided by the motion planner. Results demonstrate that the proposed motion planning and control strategy managed to decrease the lateral error with respect to the avoidance trajectory data and maximum reference lateral motion of up to 77% and 73% respectively compared to base-type PF. The proposed strategy is then validated on an actual steering wheel system through the hardware in loop test. © 2024 Elsevier Inc. All rights reserved.