The Preview Curvature Controller for a Passenger Car Steering Robot

Abstract

Purpose – Saab Automobile AB is planning to extend the use of steering robots to increase the repeatability and effectiveness of vehicle dynamics testing. A robot from Vehico GmbH has been considered for this task and to extend the use of the robot, from today’s open loop control, a closed loop path controller has been developed in this master thesis. The new controller uses GPS and inertia sensors to guide the vehicle along a predefined path. The controller can also perform superimposed closed loop control of open loop manoeuvres to enhance safety during testing. Procedure – After an initial literature review the controller was developed as a Simulink model. IPG CarMaker was used to simulate the vehicle dynamics and evaluate the performance of the controller. When the controller had been developed in Simulink it was converted to C-code and implemented into the robot software. Physical vehicle testing was performed to further evaluate performance and correlate simulations and reality. Results – The developed controller projects a preview point onto the track ahead of the vehicle. The required path curvature to reach the preview point is calculated and used as input to the controller. A predefined map of steering angle versus curvature and vehicle speed is used to calculate the required steering angle. Calibration of the controller is done by performing a simple steady state circle manoeuvre. The path following performance is very good all the way up to the limit of adhesion. The actuation is smooth and the path error rarely exceeds 1 m even at lateral acceleration levels of 10 m/s2. Most importantly good repeatability between the runs is achieved.