Fast Digital Self-Interference Cancellation based on Adaptive Momentum Gradient Descent Algorithm

In the co-frequency and co-time transceiver processing system, the transmitted signal will leak into the receiving channel to form the self-interference, resulting that the useful reception signal be distorted. In order to promote the self- interference cancellation capability in digital domain, the adaptive momentum gradient descent algorithm is proposed in this paper, which combines the exponentially decaying historical gradient and current gradient estimation together to update the weight coefficients of adaptive filter, which can reduce the impact of noise and accelerate the convergence velocity. Finally, the simulation demonstrates that convergence velocity of the proposed algorithm can be speeded up more than 4 times faster than the traditional stochastic gradient descent algorithm, and the self-interference cancellation capability can be up to 51 dB when ADC is 10 bits.