| Mobile robot technology,which integrates control,communication,perception,machinery and so on,has been a highly concerned research field.Along with the rapid development of science and technology,mobile robot technology is becoming more and more mature and gradually applied to people’s daily life,such as automated guided vehicle,welcome robot,auxiliary driving vehicle and so on.The more complex application scenarios,the higher the intelligent requirements for mobile robots.Autonomous navigation,obstacle avoidance and track tracking are the most basic and critical functions of mobile robots,which have very important research significance.This paper will study the two-dimensional map construction,path planning and navigation,trajectory tracking control of mobile robot based on lidar,and use Matlab and ROS to simulate and verify the simultaneous localization and mapping(SLAM),global path planning,local path planning and trajectory tracking control of the whole robot system.The main research work of this paper can be divided as follows.Firstly,according to the specific functions of mobile robot research,the design and construction of the hardware framework and software system of the mobile robot platform adopted in this paper are described,and a complete mobile robot hardware and software platform is successfully built.Then the basic principle of location mapping of mobile robot is studied,and the raster map building function of mobile robot is realized based on ROS.What is more,the A*global path planning algorithm and dynamic window approach(DWA)local path planning algorithm in the autonomous navigation of the mobile robot are studied and analyzed,and the autonomous navigation and dynamic obstacle avoidance function of the mobile robot are realized.Secondly,in order to solve the linear time-varying equation problem obtained from the trajectory tracking problem,the online solving problem of time-varying linear equation is analyzed and modeled,and a power-type varying gain discrete-time recurrent neural network(PVG-DTRNN)model with time-varying gain function g(t)=tp+p is proposed.The proposed discrete neural network model with variable gain has three kinds different models in three different cases.With time-varying gain parameters,the model has better convergence performance when solving time-varying linear equation problems.Finally,through Matlab simulation,it is successfully verified that PVG-DTRNN has good convergence.Finally,based on the kinematics model and geometric relationship of the robot,the trajectory tracking pose error model of the mobile robot is established,and the model predict model(MPC)is used as the virtual controller to transform the tracking problem into the quadratic programming problem.To solve the obtained problem,based on the neural dynamics method,a power-type varying gain recurrent neural network(PVG-RNN)is designed to solve the problem,and the trajectory tracking simulation is conducted in the Matlab and the experiment is performed on designed mobile robot platform,which is successfully proved that the designed PVG-RNN can track the given trajectory quickly and stably. |
