A New Method of Clothoid-Based Path Planning Algorithm for Narrow Perpendicular Parking Spaces

摘要

Autonomous driving and its practical applications for mobile robotics (e.g. automotive industry) are one of the most futuristic plans of engineering industry which will change the way of using vehicles. Having a fully autonomous vehicle on public roads is not something for the short future. To bring the autonomous driving technology to real life, the very first step would be bringing the applications of autonomous driving gradually in people's routine driving style. An automated parking system for autonomous vehicles is one of these examples in real life applications. Planning an optimum path for all types of parking scenarios under a reasonable maximum running time is an important task to have a deterministic trajectory planner. Here we worked on some of the common and critical parking scenarios for narrow perpendicular parking use cases in which the autonomous vehicle needs to handle the collision-free parking maneuver with a limited number of maneuvers to park in a tight parking space. The current planner algorithms plan several unnecessary maneuvers for narrow parking slots. Our proposed method is using Clothoid curves to plan a three-goals maneuver to park the car perfectly and avoid unnecessary maneuvers. We have implemented this method for several parking scenarios and could find a three goals collision free maneuver to park for all use cases with a reasonable average performing time (~700 ms).