Combining A* and RRT* for autonomous indoor navigation of mobile robots

Course : Motion Planning

Problem Statement : To develop a new motion planning algorithm with quick replanning capabilities to succesfully navigate in indoor uncertain static environments

Application : Motion planning & autonomous robot navigation stack for indoor service robots

Approach :

• Formulated a highly-scalable complex system design for robot motion planning consisting of global planners, local planners, controllers, obstacle detection, trajectory estimation modules from scratch.

• Employed the use of A* as global planner for achieving optimality and RRT* as local planner for quick replanning in case of potential collisions.

• Added new plug-and-play feature to add new algorithms of choice for global &/or local planner which could be easily integrated.

• Increased the algorithm database by added DFS, BFS and Dijkstra

• Developed a production quality software capable of new feature-integration with ease using C++14 in Robot Operating System with Git for CI.