Continuous autonomous driving based on computer vision and convolutional neural networks

Wei Xiaoyu, Zheng Jingyi

The article provides an overview of the main approaches to the organization of continuous autonomous driving. Particular attention is paid to computer vision and deep learning technologies. The logic of organizing the infrastructure for driving cars of high autonomy is considered. The paper proposes a new approach to the construction of neural networks based on structural and functional analysis. A software solution to the problem of continuous steering angle prediction for autonomous driving in Python based on convolutional neural networks is described.

Keywords:unmanned vehicle, autonomous driving, telematics, artificial intelligence, artificial neural networks, segmented images, hypergraphs, multi-agent system, rational agent, computer vision, pattern recognition, smart city, road infrastructure architecture, V2V-, V2I-, V2X-interaction, odometry, telematics, HAD systems, CAN bus.