Applications of mobile robots in industries, business and human living environment have been growing significantly in recent years. And mobile robots can carry out those tasks such as nuclear plants, rescue operations and other situations which are dangerous to human body. 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. So the path planning problem is defined as follows. Given a robot and a description of environment, try to plan a path between two or more specific given locations. In addition, the environment is usually uncertain as complete and the obstacles in this environment may not be known priori.

In the real sp...[ Read more ]

Applications of mobile robots in industries, business and human living environment have been growing significantly in recent years. And mobile robots can carry out those tasks such as nuclear plants, rescue operations and other situations which are dangerous to human body. 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. So the path planning problem is defined as follows. Given a robot and a description of environment, try to plan a path between two or more specific given locations. In addition, the environment is usually uncertain as complete and the obstacles in this environment may not be known priori.

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. Such as passing a series of waypoints in the set of position or orientations, hitting the target point in a certain velocity and minimizing the path length or minimizing the whole thrust.

In this thesis, I give an introduction of several different kinds of path planning algorithms, make a comparison among those different algorithms in different application situations and propose a novel hybrid method to improve the efficiency and reasonability of path planning by combing with artificial potential field method and random particle optimization algorithm. 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 track 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.