Blind Multi-user Detection Based on BP Neural Network

Combat multiple access interference(MAI) and near-far effect, to some degree, BP neural network can be better suppression interference, because of its use of the least squares method, prone to local minimum. In this paper, instead of using the MMSE algorithm for least squares and the addition of a new kind of constraints, and gain iterative formula of BP neural network blind multiuser detection algorithm, to achieve the blind multi-user detection, Simulation results indicate: with constraints BP-MMSE algorithm has better convergence properties and lower bit-error ration, to some extent, Certain limits on the effective suppression of multiple access interference, also has an effective capability against near-far effect.