Extended Target Tracking Algorithm Based Onairborne Condition

Target tracking is to predict the moving status informationby using the obtained target datawhich is widely used in the military,air traffic control,coastal surveillance and other fields.Airborne tracking uses the carrier aircraft as a tracking platform.It can be adjustedwhen the target makes a maneuver response,which is widely used in military target tracking.When the resolution unit of the observer covers lots of targets because of thelarge size,each target would generate multiple measurements at the same time.Targets that generate multiple measurements at the same time are called extended targetswhichhas become a hot topic for tracking research at home and abroad.The common target tracking methods are Kalman algorithm and IMM multi-model algorithm.These methods are relatively mature,but they all process the point target data to estimate the target motion state and cannot deal with the extended target directly.At the same time,the algorithm has poor adaptability andmany settings are already made artificiallybefore the target state is obtained.To solve these problems,the PHD probability density hypothesis filter is used to process the extended target and the IMM algorithm is used to track the centroid of the extended target in this dissertation.The main work of this study is:Analyzing the difficulties of target tracking under airborne conditions,we can solve the problem of maneuverability by using airborne coordinate transformation and suitable tracking algorithm.In order to reduce the mobility of the observation platform and the tracking target by the airborne coordinate conversion link,this dissertation converts the coordinates into the NED coordinate system required by using the information provided by the carrier.If the Calman algorithm is used in tracking algorithm selection,it will lead to large errors.The traditional IMM algorithm can not meet the requirement of high precision tracking because of its poor adaptability.By comparing the error of IMM algorithm under different transfer probability matrices of Markoff model,the multiple suboptimal fading factor is introduced to improve the adaptability of IMM algorithm.By comparing the error of the tracking target with the adaptive IMM algorithm and the traditional IMM algorithm,the adaptive IMM algorithm has a better tracking performance in the maneuvering target tracking.Analyzing the difficulties of extended target tracking,we can use extended target to be treated as point target and then point target tracking method to track it.This dissertation put forward a method to reclassify the extended target and then extract the centroid combing the SNN similarity division after obtaining the extended target number,which improve the tracking precision of the extended target.The PHD-SNN repartition method has better tracking performance when dealing with the intersection and derivation of extended targets.The radar tracking system is embedded in radar tracking platform designed by myself,and actual data is tracked and displayed.The experimental results show that the whole algorithm proposed in this dissertation can have good tracking performance for the extended target under the recorded conditions.