Control Design For Nustag Self Driving Car Project

Abstract

Automation technologies are bringing revolutionary changes in the world. Autonomous driving is subject to extensive academic research and heavy investments in the corporate sector. Reliable autonomous urban driving hinges upon vehicle control system which enable autonomous actuation of the vehicle. The objective of this project is to design the longitudinal and lateral control for the prototype vehicle of NUSTAG Self Driving Car project. The research of the project is aimed at modeling, system identification and control design for lateral and longitudinal vehicle motion. The major areas of this research include: i) vehicle modeling and system identification, ii) vehicle longitudinal control, iii) vehicle lateral control. Dynamic modeling techniques on half car and a bicycle model have been utilized to model vehicle’s longitudinal and lateral performances while, electronic equations of motors are used to formulate motor actuator behavior. The low-level steering model has been constructed on a 6-bar-linkage performing Ackerman steering. System identification is done by finding vehicle and motor parameters involved in the models. The desired path to be followed by the vehicle was precomputed. PID control scheme has been extensively applied to control the low-level throttle and steering, whereas an approach for LQR control scheme is made to make a high-level controller for vehicle’s steering. Suitable results are achieved after tuning of the controllers which can tested and further tuned by practical experimentation for real-time controller application purposes.