Research And Implementation On Key Technoldgies Of Autonomous Navigation Of Mobile Robot

With the rapid development of the computer technology and control theory, the development of the mobile robot has drawn world-wide attention in recent years. As a high intelligent system, which combines environment recognition, path planning and motion control, mobile robot is of significant scientific value. Meanwhile, the broad application prospects of mobile robot in the various fields of production and life will make it have high commercial value. It is reasonable to believe that the scientific research of the mobile robot and its application for the production and life will significantly improve the efficiency of work, and profoundly influence the way human live. The autonomous navigation processes of mobile robot mainly include robot localization, object tracing, path planning, motion control and etc. Further study on these key technologies is undertaken after the specification of those basic problems:Firstly, for the robot localization, this paper focuses on the Monte Carlo method, a common localization method. Based on the theories of this method, two phases of the robot localization, measurement update and motion update, and the whole localization process are analyzed and described.Secondly, for the object tracing problem, two important filters, Kalman Filter and Particle Filter are studied, then the detailed working process and formula deduction of the two filters are given. In addition, an effective resampling algorithm, which is very useful for the resampling phase in the Partial Filter, is presented.Moreover, for the path planning, after a close scrutiny of the uniform-cost search and A*search algorithms, the latter was chosen for its better efficiency. Additionally, in consideration of the complexity of the real situations in the robot’s motion process, the dynamic path planning is used to figure out the multi-start path planning problem; finally, the gradient descend is used to smooth the acquired path. At last, for the motion control, a popular controller in control theory, PID controller is introduced to control the robot motion. The P controller is mainly used to solve the trajectory offset problem; the D controller is used to solve the overshoot problem caused by P controller, and the I controller is used to reduce the system bias. By using these three controllers combatively, the cross-track error can be controlled in a small range. Additionally, a twiddle algorithm is applied to find out the optimal parameters of the controller.Based on above researches, Python language is chosen to implement these technologies, the robot’s environment is simulated by the Python runtime environment. And the experiment results are obtained. Practices have proven that several important issues in autonomous navigation process of mobile robot can be solved, therefore, further researches about these technologies will play a key role in this filed.