Neural network interference canceller in DS/SSMA communications with impulse noise
S.H. Leung, Jian Weng, Guoan Bi · 2002
A new neural network interference canceller for the receptions of DS/SSMA communications with impulse noise is presented. In this canceller, a nonlinearity is used to limit the large impulse noise and an Armijo gradient algorithm applied in estimating and subtracting the interfering signal. It is shown that without knowing the signals' amplitudes and the data bits a priori, the neural network interference canceller can jointly suppress the multiple-access interference and the impulse noise and provide a near single-user performance.