Obstacle-avoidance algorithm design for autonomous vehicles considering driver subjective feelings. (March 2021)

• Record Type: Journal Article
• Authors: Deng, Weishun; Guo, Huan; Zhang, Kaijiong; Lin, Mengqi; Zhang, Xi; Yu, Fan
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 235:Number 4(2021)
• Page Start: 945
• Page End: 960
• Publication Date: 2021-03
• Subjects: Autonomous vehicles -- active safety control -- artificial potential field -- model predictive control -- local obstacle avoidance; Mechanical engineering -- Congresses; Transportation engineering -- Congresses; 629.2
• DOI: 10.1177/0954407020977110
• Languages: English
• ISSNs: 0954-4070

Abstract

Normally, autonomous vehicles (AVs) are limited to be widely used in market not only for technical factors, but also psychological reasons. Considering the psychological feelings of drivers during switching manned to unmanned operation modes, an algorithm for avoiding obstacles is designed for AVs by considering driver psychological feelings. A so-called confidence-limit-distance (denoted as CLD) for driver to avoid obstacle is experimentally obtained by a number of real track tests with 100 volunteer test drivers as required to approach the obstacle in a certain way. Based on Artificial Potential Field (APF) method, a road potential field is established accordingly to characterize the information on the real road. To express the different influences of obstacles on the driver's psychological feeling in both longitudinal and lateral directions, a confidence potential field also is established based on a two-dimensional normal distribution combining von Mises distribution. Hence, the second-order Taylor expansions of the road potential field and the confidence potential field are firstly introduced into the cost functions for model predictive control (MPC). The corresponding MPC algorithm used here selects front-wheel steering angle as the control variable to be solved. The CLD and range of sensed vehicle motion state variables are taken as the constraints of the MPC. Co-simulations and Hardware-in-the-Loop (HIL) tests are carried out, showing the effectiveness of designed algorithm, which can be useful in the development and design for Advanced Driving Assistant System (ADAS) and AVs.