| Robot navigation is a core issue of robot technology. Classic Group Search Optimization algorithms such as GA, PSO both have outstanding performance when they are used to solve robot navigation problems. In this thesis, according to the requirement of social robot navigation problem, a new Social Group Optimization (SGO) is proposed based on Group Search Optimization algorithm. In SGO, social behavior theory and information entropy are introduced to model the behaviors both between robots and human being. And then a valuation function is established as the core function of SGO. And effectiveness and feasibility of the algorithm are verified theoretically and by experiment.At the beginning of this thesis, the background and significance of the robot navigation problem are analyzed in the first chapter as well as the relative researches. And then the theory foundation "Social Group Optimization" is proposed. By modeling various social behaviors between robots and human beings, behaviors are classified and quantified based on the robot navigation problem. In this process multi-robot environment is taken as a chaotic system and therefor information entropy theory is used for modeling and quantifying. Beyond that the effectiveness is verified theoretically. After the theory work, a simulation platform is built for both simulating and data analysis. The effectiveness and feasibility are also verified in experiment way. A conclusion of this paper is conducted in the last chapter and then the prospect of robot navigation problem is given. |
PDF Full Text Request