On the Influence of Rear Axle Steering and Modeling Depth on a Model Based Racing Line Generation for Autonomous Racing

Abstract

Autonomous driving at the limits of the vehicle dynamics is a challenging task, which exploits the capabilities of the vehicle. In autonomous racing the superior goal is to reduce the time necessary to drive a lap. Using rear axle steering gives additional actuation to enhance the vehicle's capabilities up to the tire friction limits and to further reduce the laptime. This paper presents a comparison of a production car to a car with an additional rear axle steering on the basis of a racing line optimization and the resulting laptime. The benefits of this additional actuation are outlined, coming from an increasing vehicle dynamics potential and from an adaptive placement on the track. Different models are taken into account during racing line optimization to compare them and discuss their influence on the results. It is shown, that a simplified vehicle dynamics model gives a suitable approximation of the optimal racing line compared to a simple lumped mass model. Furthermore, the rear axle steered car is less sensitive on modeling simplifications in comparison to the standard car.