Implementing Voronoi-based Guided Hybrid A* in Global Path Planning for Autonomous Vehicles

Planning a global path to navigate autonomous vehicles from a generic perspective defines the overall maneuvers and performance of autonomous vehicles. Inefficient and time-consuming approaches limit the performance of autonomous vehicles in planning a path to reach the desired target position. This paper presents a low-cost and computationally efficient approach of fusing the well-known Hybrid A* search algorithm with Voronoi diagram path planning to find the shortest possible non-holonomic route in a hybrid (continuous-discrete) environment for autonomous vehicles in valet parking applications. The primary novelty of our method stems from two points: at first, Voronoi diagram is exerted to introduce an improved and application-aware waypoint creator to produce the correct waypoints for the Hybrid A* algorithm and then the derived shortest optimum path regarding the non-holonomic constraints of the urban vehicles is planned using Hybrid A* search algorithm. The method has been extensively tested and validated and proven to up to 45% (30% on average) faster than the basic Hybrid A* algorithm.