Driving behavior correction in human-vehicle interaction based on Dynamic Window Approach

Abstract

This paper presents the approach of an applicable safety driving methodology for human drivers with focus on human-vehicle interaction. The approach is based on Dynamic Window Approach (DWA) in co-operation with perception of the obstacles. The human driving behaviors are modelled for the design of controller, refined by referential paths using evasive trajectory model, where linear and angular velocities are limited and corrected by DWA which performed as an obstacle avoidance strategy. Results of trajectory following and obstacle avoidance are compared with the Visual Servoing (VS) controller as a corresponding approach of autonomous control pattern.