Research On Autonomous Navigation Of Mobile Robot Based On Rough Sets Theory

Mobile robot autonomous navigation theory and techniques are still important questions researchers focus on. According to a survey of mobile robot navigation research, the idea to study mobile robot navigation by the view of uncertainty, autonomous and knowledge learning is put foreword. Rough sets theory is introduced to get robot navigation knowledge and a mobile robot navigation frame is presented. According to the requirement under dynamic structured environment of research projects, the main purpose of this thesis is to accomplish an autonomous navigation experiment system for mobile robot on which to study environment perception, behavior planning, motion control, knowledge learning and multi-robot cooperation strategy and its algorithm. Works are depicted in detail as follow:Firstly, navigation model based on rough sets theory is proposed. With the analysis of basic rough sets theory, including equivalence relation, definition of rough set and approximation, reduction and getting core principle, a navigation model is put foreword. Then, main tasks to utilize this model are given and analyzed. With the introduction of rough sets to robot navigation subject, a new robot navigation model with learning ability is proposed.Secondly, navigation knowledge acquisition method based on rough sets theory is studied. Dynamic reduction problem is discussed and algorithm adding and reducing rules in the decision-making table dynamically are put foreword so as to reduce the computing complexity. A method to get the initial rule sets by the way of finite-state-machine-rough-sets is brought up to solve the special key problem in navigation knowledge acquiring. Based on this method, mobile robot navigation experiment under multi-dynamic objects environment with single CCD is achieved and the validity of this method is proved.Thirdly, ways to improve the ability for mobile robot to percept environment are also discussed. An uncertain model for mobile robot to percept environment is put foreword and Two Iso-Vision System (TIVS) is designed. Furthermore, TIVS localization model, two vision information fusion algorithm and...