A Robot Simulation of the Tower of Hanoi Puzzle using OpenRAVE with Sensor Feedback

Abstract

In recent years, more and more efforts have been put on creating robots that operate in domestic environments. A robotic simulation environment for testing, developing and deploying motion planning is extremely necessary for both research and real-world robotics applications. In this thesis, a newly released robotic real-time simulations software library called Open Robotics Automation Virtual Environment (OpenRAVE) has been used to simulate a robot playing the Tower of Hanoi using sensor feedback. The research demonstrates and evaluates the robotic simulator OpenRAVE with sensor feedback using the Tower of Hanoi puzzle as an example to develop test scenes and programs. The algorithm developed in this thesis can solve the puzzle with different initial disks positions so that multiple simulation test scenes could be created to thoroughly and comprehensively measure the performance of robot simulator. The manipulation planning algorithms of OpenRAVE have also been discussed. The physics engine of the simulator ODE provides a real-time robot simulation with gravity and friction force applied. Four the robotic simulation environment scenes have been built in OpenRAVE and the robot has successfully accomplished the Tower of Hanoi puzzle in different circumstances. All the simulation experiments prove that OpenRAVE is a powerful and easy-to-use robot simulator with strong reliability and suitability for this research.