Publication

Path Control Algorithms for Autonomous Steering and Braking of Heavy Vehicles

In this study, path control algorithms for autonomous steering and braking interventions intended for severe rear-end collision avoidance manoeuvres of heavy vehicles are evaluated with computer simulations. First of all, in order to acquire basic kinematical data for the driving conditions prior to an accident, an accident database search was carried out using the ETAC (European Truck Accident Causation) database with the focus on three variants of rear-end collisions (due to a stopped vehicle in front, a slowing vehicle in front and a slow vehicle in front travelling at constant speed) for the thesis work. The travel speeds for the host and the target vehicles before the crash as well as basic road details were obtained. Because of simple nature of the rear-end collisions due to stopped vehicle in front, this was chosen as the accident of interest for this study. Secondly, different possible closed-loop path control algorithms utilising either steering or differential braking were designed. The steering controllers proposed in this study are either proportional (P) or proportional-derivative (PD) type control. For the differential brake control, only a proportional controller was proposed. All these controllers were tailored to include what is called prediction distance and they were integrated for possible alignments of the truck with respect to the desired escape path. Thirdly, all proposals in control algorithms were tested with a four degree-of-freedom truck model using numerical computer simulations. For development purposes, a severe path resembling the escape path of the accidents scenario of interest was designed. The effect of control parameters and the prediction distance on the path following performance were evaluated. The implementation of a prediction distance was found to help the control algorithm to keep the truck on the desired path with lesser lateral deviations and better stability. Finally, the entire set of control parameters and the prediction distance that had been found to perform well were combined and tested for a realistic collision avoidance manoeuvre which was almost in the absolute limit of wheel lift-off for the heavy vehicle. It was seen that reduction of speed was necessary for the vehicle to be able to follow the path with minimal lateral deviations and without wheel lift-off. For this reason, open-loop application of soft service braking throughout the manoeuvre was investigated and it was found that path control performance had increased.

Author(s)
ÖDÜL BIRGE BILEN
Research area
Systems for accident prevention and AD
Publication type
Master's thesis
Year of publication
2010