Enhancing Robot Navigation in ROS 2 with Theta* Planning

Path planning is an essential aspect of autonomous navigation as it greatly affects robot movement in dynamic, constrained environments in terms of efficiency and feasibility. This paper presents a comparative analysis of Theta* and A* algorithms applied in the simulation framework of ROS 2 and Gazebo. In assessing the methods in a structured environment, both algorithms were separately tested under identical conditions in terms of their efficiency in computing an optimal path, given traversal constraints. While the original A* algorithm computes paths by considering points, the Theta* algorithm has been extended to optimize line-of-sight paths, eliminating unnecessary waypoints and creating smoother traverse alternatives. On the contrary, A* computes paths in a strictly grid-based manner, guaranteeing feasible yet often sub-optimal paths, with more turns. The experimental findings, analyzed using both qualitative and quantitative perspectives, thus show the implications of these algorithms on path smoothening, computational efficiency, and obstacle avoidance. This work lends insight into determining the real-time path-planning efficacy in robotic applications with a view towards selecting an appropriate algorithm for distinct autonomous navigation tasks.