Artificial Moment Method For The Local Path Planning Of Mobile Robots

The local path planning of mobile robots is an essential and challengeable problem in robotics. For the problem, the existing algorithms either cannot meet the requirement of the real time property or often yield low quality solutions. Because a new robot model, that is, each robot has a Principal Motion Direction line (PMDline), is used in the artificial moment method for multi-robot formation control, the method shows many attractive features in the motion control of mobile robots. This paper extends the existing artificial moment method to the local path planning of mobile robots in complex dynamical environments.Firstly, for the single robot path planning problem, connected function segments are proposed and rules for setting N-type and B-type connected function segments are given respectively to prevent robots from walking into too narrow regions and to decrease the number of knowledge obstacle walls. An idea based on attractive segments for attractive points is proposed and three conditions are given for new attractive segments. An algorithm is presented for attractive segments based on close and lateral tangent points. Attractive moments, repulsive moments and an artificial moment motion controller are designed. A tangent point artificial moment method is proposed for the local path planning of a single robot, and the general steps of the method are given.Secondly, for the disadvantages of the tangent point artificial moment method, that is, the computational burden is too heavy and the performace cannot be guaranteed in some complex environments, a new algorithm for attractive segments is presented based on key obstacle function segment sets. The method for key obstacle function segment sets and a new rule for5-type connected function segments are also presented. An obstacle function semgment artificial moment method is presented for the local path planning of single robot, which performs better in some complex environments. The general steps of the method are given.Finally, for the multi-robot path planning, the methods for attractive segments and connected function segments of the obstacle function segment artificial moment method are extended to multi-robots. An idea by using attractive angles to determine robots’PMDline directions is proposed and an algorithm for computing attractive angles and for improving attractive points is presented. Coordinated moment is proposed for resolving the conflicts between robots. New attractive and repulsive moments and artificial moment motion controller are designed. As a result, an artificial moment method for multi-robot path planning in complex dynamical environments is obtained.Some properties of the proposed artificial moment methods, such as global convergence and safety, are also discussed, and several theorems about the properties are presented. Extensive simulations are given and the results indicate that the proposed methods are feasible.