Enhancing the PRM Algorithm: an Innovative Method to Overcome Obstacles in Self-Driving Car Path Planning
Mohamed Bakir, My Abdelkader Youssefi, Rachid Dakir, Mouna El Wafi · International Review of Automatic Control (IREACO) · 2024
Navigating environments quickly while trying to minimize the time taken and the deviation from the path is a challenge for self-driving cars especially when dealing with narrow passages. The Probabilistic Roadmap (PRM) algorithm has been used to address this challenge in vehicles and route planning. Its effectiveness in handling narrow passages is still a key concern. This article describes a method of combining sampling methods. One of these is random sampling, while the other one is Gaussian sampling. Both are based on PRM. However, the objective is to assess the value of this approach when exploring terms that are difficult for PRM-based approaches. The testing of this method in autonomous vehicle navigation that is done in the real world provides information about its effectiveness specifically when maneuvering through spaces. By combining multiple sampling methods with the PRM algorithm, this method attempts to address the issues associated with passages and increase the flexibility of path planning algorithms. While it is noted that this method has advantages like more dynamic maneuvering and a lower amount of distance to cover as well as the walking path, the essence here is on whether it can face obstacles in the passages. This approach appears encouraging in reducing the time consumed while driving an autonomous vehicle as well as increasing their navigation efficiency in intricate settings.