This paper presents a framework for tracking\na mobile ground target (MGT) using a fixed-wing\nunmanned aerial vehicle (UAV). Challenges from pure\ntheories to practical applications, including varying\nillumination, computational limits and a lack of clarity are\nconsidered. The procedure consists of four steps, namely:\ntarget detection, target localization, states estimation and\nUAV guidance. Firstly, the MGT in the wild is separated\nfrom the background using a Laplacian operator-based\nmethod. Next, the MGT is located by performing\ncoordinate transformations with the assumption that the\naltitude of the ground is invariant and known. Afterwards,\na Kalman filter is used to estimate the location and velocity\nof the MGT. Finally, a modified guidance law is developed\nto guide the UAV to circle and track the MGT. The\nperformance of our framework is validated by simulations\nand a number of actual flight tests. The results indicate\nthat the framework is effective and of low computational\ncomplexity, and in particular our modified guidance law\ncan reduce the error of the tracking distance by about 75%\nin specified situations. With the proposed framework,\nsuch challenges caused by the actual system can be tackled\neffectively, and the fixed-wing UAV can track the MGT\nstably.
Loading....