Modeling and Control of UAVs using Neural Networks

*† ‡ § ** †† , This paper presents the research being done at Cal Poly Pomona on the dynamic modeling and control of unmanned aerial vehicles (UAVs) using neural networks. The overall goal of the ongoing project is to develop and validate the neural network-based robust nonlinear controllers using a number of UAVs. Another goal is to compare the performance of the controllers designed using various methods. This paper talks about the development of custom avionics system for the UAVs, flight data collection, dynamic modeling, and training of networks using Radial Basis Function and Multi-Layer Perceptron networks. Simulation results are shown. Nomenclature p = airplane roll rate, rad/sec q = airplane pitch rate, rad/sec r = airplane yaw rate, rad/sec V = freestream velocity, ft/sec x F = force along X axis, lb y F = force along Y axis, lb z F = force along Z axis, lb g = acceleration due to gravity, ft/sec 2