ITE-RRT*: Intelligent Path Planning for Autonomous Cars With Intermediary Trees, Triangle Inequality, and Equal Distance Optimization

Abstract

This research introduces a new algorithm that enhances the existing Intermediary RRT*-PSO and Informed RRT*, focusing on enhancing the path’s feasibility by replacing edges around the corners with trapezoidal-based turns to ensure robot’s safety and functionality. The algorithm begins by generating an initial path using Intermediary RRT* algorithm. This path is then further refined using an informed sampling strategy, while being optimized by the Equal Distance optimization scheme. The proposed algorithm not only finds the initial path faster than previous methods, but also creates a solution with reduced sharp edges around corners. Non-holonomic robots benefit specifically when using this method because it reduces the frequency of abrupt turns. The procedure was tested through simulated runs on changing maps to establish both solution-generation duration and path-length requirements. This method is shown to be effective for map-based navigation across different terrains by delivering improved convergence rates and reduced path lengths over earlier navigation techniques.