Automation Of Vehicle For NUSTAG Self-Driving Car

Abstract

Automation technologies are bringing revolutionary changes in the world. Autonomous driving is subject to high academic research as well as high investments from the corporate sector as it can reshape the transportation system. Reliable autonomous urban driving hinges upon vehicle’s ability to perceive and navigate the environment which will enable improved driver and vehicle safety, driving efficiency, driver assistance, traffic management and other field robotic applications. The objective of this report is to describe the automation systems designed for the prototype vehicle of NUSTAG Self-Driving Car project. The research of the project is aimed at modeling, design, system identification, and implementation of autonomous technologies. The major areas of this research include: i) vehicle controls design and implementation, ii) electric motor control and steering actuation, iii) state estimation, iv) perception v) path planning. Self-Driving cars are poised for accelerated adoption, with an ability to create and keep a map of their environment based on the input form the variety of sensors located at multiple parts of the vehicle. The software then processes all these sensory inputs, determining the instructions to be sent for the vehicle’s actuation. Vehicle uses visual inertial odometry for state estimation, convolutional neural networks and computer vision for perception and stereo vision-based 3D object detection for localization of obstacles, along with path planning algorithms to plan the local and global path, which is followed by the actuation of low-level controller. All these cutting-edge technologies enable efficient autonomous driving.