Note: development for this project was haulted in November 2020 to respect my NDA with my employer.
This is a fork of the original SpotMicro repository
This was done for the ACS Master's Project of GVSU The following project is an extension of the simulation for the quadruped robot “SpotMicro”, inspired by the “Spotmini” of Boston Dynamics. The simulation is run in PyBullet, a Python-based 3D physics simulator. The code base for the simulation has not been updated in several years and only contained basic functionality and a sample reinforcement learning algorithm. The first part of the project was mostly pertaining to learning the source code and PyBullet, and then generating a maze based on a 2D image. Several variations of the maze, with varying levels of complexity, were used for the simulation, with the final one being a rather simple one for ease of demonstration and testing purposes. The second part of the project was manipulating the existing navigation logic of the quadruped robot to pivot when colliding against obstacles, in this case the maze walls, to ultimately find its way out of the maze. Ultimately, the robot was able to reach its goal of finding an exit from the maze. The same logic should be transferrable to a real-world build and be applicable for lager mazes and/or scenarios which can be potential future areas of exploration.