Adaptive Inverse Power Control Using an FxLMS Algorithm

In this paper, a novel adaptive inverse power control method that is based on the filtered-x LMS (FxLMS) algorithm is introduced. Inverse power control minimizes the interference that a cognitive radio creates to licensed users and allows more users to share the spectrum. It is also needed in multi-access systems like code division multiple access (CDMA) to alleviate the near-far problem. The proposed variable step algorithm adjusts the step size in a nearly optimal way. Based on numerical analysis this new method clearly improves system performance compared to the algorithm where the well-known fixed or variable step adjustment power control is used. A normalized version of the FxLMS algorithm is needed in a fading channel. In a slowly fading channel (e.g., normalized Doppler frequency of 0.001) the FxLMS power control can keep the received signal-to-noise ratio (SNR) in the desired level with a good accuracy most of the time. The standard deviation of the received SNR is 1.92 dB when the received SNR is kept at 10 dB. The results with almost all different SNR values are better than the other methods can achieve.