Examination Committee

Prof Ling SHI, ECE/HKUST (Chairperson)
Prof Wai Ho MOW, ECE/HKUST (Thesis Supervisor)
Prof Kam Tim WOO, ECE/HKUST
Prof Fu ZHANG, ECE/HKUST

Abstract

Applications of mobile robots in industries, business and human living environment have been growing significantly in recent years. In order to realize the robots’ assisting and replacing of human’s work in some missions, many challenging problems in mobile robots research need to be figured out, the important one of them is path planning.

In the real specific applications, collision-free property of path for the robots’ movement is the fundamental requirement. A number of algorithms have been developed for that fundamental requirement and other further optimization criteria of the path planning.

In this thesis, a novel hybrid approach for path planning of mobile robot in indoor complicated environment is introduced. This method is proposed to improve the efficiency and reasonability of generating collision-free and smooth trajectory by combining with artificial potential field method and random particle optimization algorithm. A comparison among the proposed hybrid approach and those traditional algorithms is made and generated trajectories results are presented. Besides that, I also present specific application situations of this hybrid algorithm when an original offline trajectory has been given and many optimization criteria are made. I introduce that concrete algorithm structure of implementing proposed hybrid method to avoid obstacles when the robot tracks original minimum snap trajectory. And dynamic partially minimum snap trajectory generation method is designed which have a good performance on trading off saving time on obstacles avoidance and smoothing trajectory in dense dynamic environment. A variety of simulation experiments in different situations have been conducted to verify the validity of those proposed algorithms.