Research On Composite Navigation Technology Of AGV For Material Distribution

With the year-on-year development of science and technology,intelligent robots have penetrated into all areas of people’s daily lives and are gradually replacing manual labour.Automated Guided Vehicles(AGVs)are especially widely used in industrial transport,featuring good flexibility,high levels of automation and intelligence,and are commonly used in workshops to optimise production patterns and promote the upgrading of production transport chains.At the same time,new processes are increasing the requirements for the navigation process and positioning accuracy of AGVs for material distribution,and how to improve the accuracy to meet the process requirements is becoming the focus of research.In this paper,laser SLAM is used as the navigation method to study the map building and fusion positioning,path planning and tracking of AGVs:(1)Based on the analysis of the current situation of AGV research and development at home and abroad,this paper uses the Ackermann steering model cart as the hardware platform.In order to meet the demand of material distribution,the external structure of the body and the hardware equipment of each module sensor are selected,while the STM32 as the core of the main control circuit board is developed and designed.(2)Establishing a good a priori map is the key to navigation operation.By simulating and analysing both Gmapping and Cartographer in different scenarios,the Cartographer algorithm,which is most suitable for the current experimental environment,was selected as the a priori map for path planning,taking hardware processing capability,map size and AGV movement speed as preconditions.The a priori map was selected as the path planning map.To address the situation that the use of a single sensor has a large measurement accuracy error,the traceless Kalman filter(UKF)is used to fuse IMU-odometer data and finally combined with the adaptive Monte Carlo particle filter(AMCL)algorithm to improve the positioning accuracy of the AGV at the global level.(3)Path tracking simulation verification analysis.In the Matlab simulation platform,the path tracking controller(MPC)obtains the information of the reference planning path and the state information of the vehicle,and then optimises the front wheel turning angle of the vehicle as the control input to realise the path tracking control of the vehicle,while designing the obstacle avoidance conditions,simulating and analysing the tracking effect of the MPC and PID controller,and experimentally verifying the control performance of the control algorithm,proving that MPC has good effect on path tracking.(4)Build an experimental platform for real vehicle verification and analysis.By developing a communication and scheduling platform,data intercommunication between the upper computer side and the ROS side of the cart can be realized.The MPC controller was also applied to the Ackermann front wheel steering trolley.The experimental results showed that the AGV could continuously avoid obstacles along the planned path to reach the end of material distribution,and the tracking trajectory was highly overlapping with the reference route and the overall trend was similar,which proved the effectiveness of MPC in path tracking.