Research On Indoor Positioning And Trajectory Tracking Motion Control Technology Of Sorting AGV

The sorting AGV is a safe and protected transport vehicle that can be guided by electric power,magnetic force or other devices to follow a pre-planned trajectory in an automated warehouse.In recent years,more and more people tend to buy goods on the Internet,especially in the "double a" year "618" and so on electric business promotion festival,express order number increased year by year,brought great pressure to the logistics industry,and sorting AGV as a kind of typical logistics transportation equipment,improve the work efficiency and logistics service ability become to realize intelligent logistics,the key to alleviate the pressure of the logistics.In order to realize the popularization of small and medium-sized automatic warehouse sorting AGV application,it is of great significance to study a low-cost,high-precision AGV navigation system.The research contents of this paper mainly include the establishment of the global coordinate system based on the AGV global coordinates and the local coordinate system based on the dolly body,the establishment of the AGV kinematic model based on the dolly kinematic relationship,and the determination of the control quantity of the AGV motion control.The positioning model based on photoelectric encoder,inertial measurement unit and ultrasonic sensor is established,and the real-time positioning algorithm of corresponding positioning model is obtained.The influence of the arrangement of ultrasonic sensors on the measurement error of ultrasonic sensors is analyzed.By adopting the feature level information fusion technology,the double-layer Kalman filter algorithm is designed to fuse the positioning information measured by the navigation calculation method,inertial navigation positioning method and ultrasonic sensor,so as to avoid the random errors in the measurement process of each sensor and greatly improve the POSITIONING accuracy of AGV.In view of the fact that ultrasonic sensors are prone to location failure due to occlusion in storage environment,radial basis function neural network is used to calculate the measured location information of ultrasonic sensors,and the nearest neighbor clustering learning algorithm is used to complete the online learning of neural network.In MATLAB programming files to complete the design of the algorithm simulation.The AGV trajectory tracking process was analyzed,and a mathematical model was established based on the deviation between the actual AGV pose and the given trajectory target.The sliding mode variable structure controller is introduced,the switching function is designed by inversion design method,the sliding mode motion exponential approach law is improved,the chattering phenomenon is weakened,the simulation model is built in MATLAB,the program file is written to complete the controller parameter setting and simulation,and good simulation results are obtained.The transfer function of each link of dc servo system is calculated and the current loop speed loop double closed loop controller is designed to improve the dynamic response ability of DC servo system.