Résumé:
Autonomous Mobile Robots (AMR) are robotic systems capable of navigating in
environments without human intervention. Their growing popularity and practical
applications have led to a rapid expansion, driven by increasing interest and research.
However, a major challenge faced by these systems is the generation and execution of
movements required for efficient trajectory planning, which remains a persistent
problem in autonomous systems. In this study, our objective is to contribute to the
field of motion planning by introducing two new variants of the Rapidly-exploring
Random Tree Star (RRT*) algorithm that integrate the Whale Optimization Algorithm
(WOA) to generate near-optimal trajectories. To validate the proposed variants, we
implemented them in a simulation environment. Then, we explored the parameter
space of WOA for both variants in order to identify optimal parameters and deepen
our understanding of behavior with different configurations. The results obtained from
the two variants demonstrate significant improvements in trajectory quality,
surpassing the performance of the original RRT* algorithm. These promising results
highlight the untapped potential of using this optimization technique and also pave the
way for further research to explore and exploit the benefits of parallelization aiming to
enhance the efficiency and effectiveness of these variants.
