Path Planning for Autonomous Cars

Abstract— Path planning plays a vital role in autonomous driving. It is the replication of the reasoning and decision-making of a human brain. This paper is about analyzing and optimizing a GitHub project which is related to path planning for autonomous cars. This work has fixed the program to drive a car on the simulated highway while avoiding collision, following traffic, safely changing lanes, and minimizing jerk. Additionally, more critical scenarios have been identified to make driving experiences safer and the present cost functions are optimized to make the car react effectively. Throughout the multiple experiments, a more efficient program has been produced that has reduced the time to finish one lap by roughly 10 seconds. Path planning is one of the extremely fundamental processes within autonomous driving. There are still challenges to make path planning safer and robust, such as behavior modelling on other cars, and more efficient path searching algorithm, etc.