D* Lite-Based Navigation Algorithm for Multiple Spherical Underwater Robots Collaboration

This research investigates the multiple robotic systems navigation in an unknown environment problem. As an attempt to solve this challenge, a navigation strategy based on D* Lite search algorithm is proposed. To ensure an optimal path to the goal, we implemented the D* Lite algorithm and improved the safety of the algorithm by defining the non-reachable cells accordingly to the size and position of the obstacles. The collaborative navigation is ensured by a fault tolerance strategy based on a low computation cost fault detection and leader switching algorithm. The performed experiments in a water tank and the simulation demonstrated that the search algorithm can safely find the shortest path to the goal. The robustness of the path planning algorithm is demonstrated further with relatively few collisions occurrence. Also, the simulation shows that the robots can safely adapt their formation from V-shape to line shape when obstacles are present in the environment. However, the formation navigation presents some inefficiencies in terms of accuracy. The developed strategy will be extended in future works with more experiments.