| In the mechanical scanning process of the machine sweeping radar,due to the beam width,a single target will form a series of points in different directions of the same distance unit during the beam's dwelling time,showing a similar “eyebrow” characteristic.The intrinsic characteristics of sweeping radar cannot be solved by hardware transformation.Therefore,this paper hopes to optimize the algorithm or cooperate with multiple algorithms to solve this feature of sweeping radar.According to the mechanical characteristics of the machine sweeping radar,the tracking system tracks the data while scanning.The system processes the received data at each wave position and completes operations such as the continuous tracking of confirmed targets,the start of new target paths,and the erasure of dim target tracks within a wave length.Real-time processing of continuous wave data enables real-time tracking of the target.According to functional requirements,the tracking system includes points aggregation,clutter identification/elimination,track initiation/confirmation,track classification,data association,track maintenance,track extinction,human-computer interaction display and other modules.This article describes in detail the algorithms used in each module,including the centroid method used in the point aggregation process,the logic method used by the track initiation/confirmation module,the probability data association algorithm used by the data association module,and the joint probability data association algorithm.And the interactive multi-modal algorithm used by the entire system architecture.Secondly,this paper presents a solution to the target tracking system based on machine-sweeping radar and implements it in detail.Finally,the human-computer interaction interface is designed,and the human-machine interaction design of the robot sweep radar is realized simply,and the correctness and practicality of the target tracking system are verified. |
PDF Full Text Request