Self-localization Of Soccer Robot Based On Quantum Immune Algorithm

With the development of science technology and the improvement of the human need,mobile robots increasingly develop towards the direction of autonomous. Robotself-localization is the basic problem to achieve robot autonomation, and also the first thing todo to achieve the navigation. It is the important and difficult point in robotics technologyresearch. The key study is to how to improve the accuracy, efficiency and stability of robotlocalization system. From the perspective of system architecture and optimal control, soccerrobot self-localization problem is studied in the competition environment.The paper research platform is based on Xi’an University of Science and Technologysoccer robot system. In the dynamic environment, the quantum immune algorithm combinedwith the localization feature points is used to improve the system of soccer robotself-localization, solve the lost state and enhance the stability and intelligent of the localizationsystem.Firstly, this paper reviews the development and research status of mobile robotself-localization, analyzes the current main research methods in this field, and describes thesoccer robot hardware structure about vision system and localization environment, and therelated localization models are established. Due to uncertainties existed in the robot soccerlocalization process, the robot pose can be obtained by the probability localization algorithm.It has been widely used, and the advantages and disadvantages of Markov,Extended Kalmanfilter and Monte Carlo localization algorithm is mainly analysed and compared. On the basis,the quantum immune algorithm for mobile robot self-localization is proposed.Secondly, for the soccer robot self-localization, using the grid method to divide the field,selecting the white feature point, and by the conversion between the robot coordinate systemand the world coordinate system, the white feature point position in the world coordinatesystem can be obtained. The initial population can be got by the immune algorithm based onthe quantum algorithm. The affinity function is used to calculate the matching degree between the white observation point and field feature point. The population are updated by the selectionoperator and mutation operator. The better individual can be earlier obtained. Hence, the robotlocalization individual can be acquired. This algorithm can ensure the diversity of populationand speed of population convergence, prevent the population premature convergence andimprove the global search capability.Finally, the robot self-localization algorithm is verified in the actual soccer competitionenvironment. The result shows that the algorithm can be applied in dynamic football gameenvironment, solve the robot lost phenomenon, enhance the effectiveness of soccer robotlocalization, reduce the effect of robot localization errors and improve soccer roboticlocalization system. The target of this paper is achieved. And our school soccer robot team hasmade outstanding achievements in national competition for recent years.