Research on Security of Key Algorithms in Intelligent Driving System

With the rapid development of the smart driving technology, the security of core smart driving algorithms is becoming more and more important. This paper illustrates the architecture of the smart driving system, analyzes the core algorithms in the smart driving system, and selects the representative local path planning and obstacle detection algorithms. By studying the local path planning algorithm based on the artificial potential field method and obstacle detection algorithm based on the V-parallax method, it analyzes the failure conditions of local path planning and obstacle detection algorithms of the smart driving system, and proposes that when the vehicle falls into the local minimum point of the potential field and when the obstacle detection algorithm is affected by the noise and threshold or there is a slight parallax change in the same obstacle, these two algorithms will lose effect, leading to misjudgment by the smart driving system. This study will lay a foundation to improve the security of the smart driving system.