Optimized RRT Planning With CMA-ES for Autonomous Navigation of Magnetic Microrobots in Complex Environments

Magnetic field-driven microrobots have shown high potential in the field of medical applications. The utilization of magnetic fields is particularly favorable due to its ability to penetrate deep tissues while ensuring high safety. Despite significant advancements in the fabrication, functionalization, and locomotion of magnetic microrobots, autonomous navigation is of paramount importance for magnetic microrobots. In light of this objective, this article introduces a novel navigation framework, using an improved path planning navigation method. The proposed method introduces a path planning algorithm, covariance matrix adaptation evolution strategy (CMA-ES) and rapidly-exploring random trees (RRT) (CMA-ES-RRT), which skillfully combines the advantages of both CMA-ES and RRT. The proposed framework not only guarantees a smooth path but also takes it a step further by significantly minimizing the overall navigational path length. These dual benefits are especially critical in medical applications, significantly improving the convenience of subsequent path tracking. Through meticulous algorithm comparisons and thorough analyses, our approach emerges as a superior choice, excelling in both path smoothness and length optimization. Extensive environmental validation analyzes unequivocally demonstrate our method's superiority over traditional RRT and its variants in terms of path smoothness and navigation path length.