Research On UAV’s Precision Attack Technology Of Ground Moving Target

Air-to-ground attack is a kind of air-ground battle military operation, and has a veryimportant application in the actual combat. In the view of the emphasis of zero loss in modern warand the high demands of efficiency, unmanned aerial vehicles (UAVs) are considered to be thebest choice as a space weapon launch platform for the precision attack. Thus, many relatedtechnologies are developed by lots of countries. By using recognition, locating and trackingmoving targets on the ground, UAVs are able to obtain the real-time status of enemy targets, whichlays a solid foundation for UAV precision attack. With the battlefield environment gettingincreasingly complex, the warfare of UAVs extends to coordinating operations avoiding theindependent operations. Therefore, multi-UAVs coordinating attack has become a hot topic in thefield of UAVs, while target assignment is a basic technique. This thesis studies the method ofrecognition, locating and tracking moving targets on the ground, and then studies satisficingallocation decision of air-to-ground target for UAV. The main contents and results are as follows:Firstly, the preprocessing is done on the obtained target image and the target recognition iscompleted on the basis of the preprocessing target image. The image preprocessing can improvethe image quality and the threshold segmentation method is studied for the target image. Based onthe segmentation results, a modified invariant moment method is developed to recognize themoving targets which play an important role for the following target locating and tracking.Secondly, a locating method is given for the ground moving target based on air-borne SARimages. The R-D equation is used to calculate the latitude and longitude of target from the radarimage pixel coordinates. Then, the target point’s coordinate under airborne coordinate system andgeographic coordinate system are obtained via the coordinate transformation. Thus, the locating ofground moving target is completed.Following, in order to obtain the tracking information, two different tracking algorithms ofparticle filter and mean shift are presented. An inherent problem with particle filter is the so-calledparticle degeneracy, i.e., how to balance the number of the particles and real time tracking is acritical issue to consider. An improved tracking algorithm is studied by combining particle filterwith mean shift. The satisfactory tracking performance is obtained.Finally, the target allocation method is studied for the multi UAVS coordinating attack. Byconsidering that the traditional distribution method costs high iterations, the satisficing decision isstudied for the target allocation. In this method, the purpose of task assignment is achieved by obtaining a satisfactory value rather than the optimal value. Furthermore, quantum geneticalgorithm is applied to optimal allocation results. The numerical simulations are given under theMATLAB environment. Simulation results show that the developed satisficing allocation decisionalgorithm is effective for the air-to-ground target of the UAV.