A terrain matching navigation algorithm for UAV

Terrain matching positioning navigation uses the traditional algorithm to linearize the terrain and compares with the datum terrain. It can provide more accurate positioning, but due to the strong nonlinear characteristics of terrain, in the process of terrain linearization, most of the terrain features are ignored. So the matching result error is relatively large. Because the pulse coupled neural network has the characteristics of neuron's characteristic of linear addition and nonlinear multiplication adjustment coupling, it does not need to establish an accurate mathematical model, and it can better deal with nonlinear problems. Therefore, this paper will explore the application of pulse coupled neural network in terrain matching positioning. Firstly, the distance difference matrix between the prior terrain data and the measured matching terrain is normalized, then the nodes on each matching surface of the topographic map will be searched and summed one by one through the coupling relationship between the similar nodes, and the magnitude of ignition times represents the similarity degree. By selecting different regions and different parameters in the topographic map for simulation, the high-precision performance of the algorithm is verified.