Minimal resource allocation network for adaptive noise cancellation

An investigation into the performance of the recently developed minimal resource allocation network (MRAN) for adaptive noise cancellation problems is presented and a comparison made with the recurrent radial basis function (RBF) network of Billings and Fung. An MRAN has the same structure as an RBF network but uses a sequential learning algorithm that adds and prunes hidden neurons as input data which are received sequentially to produce a compact network. Simulation results for nonlinear noise cancellation examples show that an MRAN, with a much smaller network, produces better noise reduction than the recurrent RBF