Path Planning of Autonomous Vehicles Based on Deep Learning Technology

Abstract

With the continuous improvement of the intelligent level of unmanned vehicles, its operating range is becoming wider and wider, the road traffic environment it faces is becoming more and more complex, and the possibility of various uncertain events is also increasing. It is of great significance to study the global path planning that adapts to the real traffic network. The purpose of this paper is to study the simulation experiment of autonomous vehicle path planning under deep learning technology. The background and significance of the research on unmanned driving strategy, the current development status of autonomous driving technology in the world, and the current development status of deep learning technology are introduced. The requirements of the control cycle duration when the car is running at high speed. The BIT* algorithm is improved, and the NBIT* algorithm is proposed, which greatly improves the planning speed when dealing with a large number of path planning tasks. The path planning of the 200 images in the NBIT* algorithm test set takes 1.61 s in total.