Computing trajectories in 3D space to avoid obstacles using systematic and stochastic search algorithms

Abstract

Recent years have been characterized by great technological advances. When it became an option to exploit in industrial work, the tasks were automated with the use of robots. In this dissertation, we present a custom-made virtual robotic arm capable of constructing Jenga towers. In contrast with previous work, we do not use force sensors for placement detection. Furthermore, to demonstrate the robot controller 's extensibility in design, we run two different simulations, ending with two different structures. Simplicity and efficiency have been considered to achieve the desired results in the experiments above, raising the key points where computational costs cannot be reduced, while providing simple alternatives for the rest of some common issues in robotics. We also compare and present results concerning various topics regarding commonly used algorithms including A*, after implementing them for the purposes of this dissertation.