Research On Tracking Of Moving Target Based On The Controlling Of Four-axis Motion Control Platform

Traditional motion control is mainly used in the field of numerical control.Most traditional motion controllers can only be used for specific objective because of their poor generality and openness.In this paper,an open motion control platform is designed and the visual inspection module is added,which form a closed-loop control system and improve the intelligence and accuracy of the platform.The visual inspection module is used to detect moving target,and the results are transferred to the four-axis motion platform to control the axes and the robot real-timely,which can realize intelligent tracking and catching.In this paper,the motion controller of the four-axis motion control platform is the double CUP structure of ARM and PCL6045 B.As the main control CPU,ARM not only has a high-speed operating frequency,but also has a wealth of peripherals,which can help to extend or crop the functions according to the requirements.PCL6045 B is a professional motion control chip,which can achieve uniaxial positioning,linear/S-curve acceleration and deceleration as well as the limit and homing and other functions.A robot is added to the four-axis motion platform as a peripheral to catch the moving target.The corresponding motion control functions are encapsulated as APIs in the software architecture of the whole platform,which can be helpful for the secondary development of the four-axis motion control platform and improve the openness of the platform.In order to realize the intelligence of the platform,visual inspection module is added to detect the moving target,and the moving target is tracked and grasped according to the real-time control of the four-axis motion control platform.The positioning and catching of stationary target need to be realized first to accomplish the accurate tracking and catching of the moving target.The purpose of the procedure is to complete the conversion between the camera coordinate and the platform coordinate.Then,the platform coordinates of the moving target are used to control the axes and robot of the platform so as to tracking and grabbing the moving target.In this paper,the TLD algorithm is used to detect the moving target.The algorithm is a single-objective long-term tracking algorithm,which combines the traditional detection algorithm and traditional tracking algorithm.The advantage of the TLD is the ability to re-detect the target in the case of occlusion and deformation of the target with high stability.The algorithm consists of tracker,detector,integrator and learner.Finally,a specific motion control scheme is proposed according to the performance of the three different kinds of motion to make the tracking of the motion platform more smooth and make the motion control system more useful.Besides,a series of experiments are designed,which verified the accuracy,real-time and stability of the whole system by tracking and grabbing moving target in different scenes.