Research On AGV Positioning Algorithm Based On Multi-sensor Fusion

In recent years,with the development of embedded,automation and artificial intelligence technology,the modern logistics industry presents the trend of digitalization and intelligence.Warehouse management is an indispensable part of modern logistics,in which the automatic guided transport vehicle is the main equipment for cargo handling in the intelligent warehouse.It is safe,efficient and labor-saving.It is an important part of modern logistics and industrial automation,and under the guidance of the trend of artificial intelligence technology,it will cause a new round of rapid development.The main contents of this paper are as follows:1.Complete the research platform of positioning control system,including hardware construction,software platform construction,multi task and priority design,sensor data acquisition and processing,etc.2.Through the research of key algorithms based on laser and vision location,combined with the advantages and disadvantages of different sensors and location algorithms,a multi-sensor simultaneous interpreting performance evaluation system based on fuzzy mathematics is established on the AGV platform.Evaluate several multi-sensor combined positioning methods and define the combination mode of multi-sensor.3.Design of multi-sensor fusion location algorithm.According to the evaluation results,a multi-sensor fusion positioning algorithm based on monocular camera,IMU and wheel odometer is proposed to solve the problems of accuracy and robustness of vision positioning system.In this algorithm,the problem of frequency synchronization is solved by data fusion between IMU pre integration theory and wheel odometer.An initialization algorithm is proposed to recover the scale information of camera and estimate and optimize the external parameters.A graph optimization framework integrating IMU,wheel odometer and camera is designed.In this framework,the external parameters are estimated online,the specific form of Jacobian matrix of optimization variables is derived,and the pose optimization is carried out by using a variety of sensor information;The fusion mode of loose coupling is designed,and the fusion framework of IMU and odometer is proposed to provide positioning function when the visual positioning system fails.Through the AGV experimental platform designed in this paper,the experiment of fusion location algorithm is carried out.The experiment shows that the fusion location algorithm proposed in this paper has better positioning effect in accuracy and robustness.