Fault detection based on RBF neural network in a missile´s actuation system

A failure observer based on RBF neural network is developed to realize the failure detection of a missile´s actuation system, and a two-level learning method for designing radial basis function (RBF) network based on improved particle swarm optimization (PSO) and regularized orthogonal least squares (ROLS). The trained RBF observer works concurrently with the actual system. By comparing the estimated output with the actual measurements, residual signal is generated and then analyzed to report the occurrence of faults. The experimental results show that the failure observer based on the RBF neural network is effective in detecting the failure of the missile´s actuation system.