engleski

Evolutionary Planner of 3D Trajectories

This paper presents a method based on evolutionary algorithms for the planning of robotic trajectories both in 2D and 3D space. Space can be arbitrary shaped containing randomly sized and distributed obstacles. Different methods were implemented and evaluated for improving the speed and the level of the convergence of the algorithm. It is shown that evolutionary algorithm in its simplest form cannot effectively solve this problem in 3D. A thorough statistical analysis is performed for various instances of the algorithm, and the best algorithmic procedures are identified, documented, and explained in the paper.

motion planning ; evolutionary algorithms ; artificial intelligence

nije evidentirano