Research on geometric algebra-based robust adaptive filtering algorithms in wireless communication systems

Abstract Noise and interference are the two most common basic problems in wireless communication systems. The noise channels has characteristics of randomness impulsivity, so performance adaptive filtering algorithms based on geometric algebra (GA) second-order statistics is greatly reduced In order to improve systems, this paper proposes novel GA-based algorithms, which deduced from robust minimum error entropy (MEE) criterion joint (MSEMEE) MEE mean square (MSE) with help GA theory. modeled by $$\alpha$$ α -stable distribution good agreement actual data paper. Simulation results show that for deviation (MSD) learning curve, (GA-MEE) algorithm faster convergence rate better steady-state accuracy compared maximum correntropy (GA-MCC) under same generalized signal-to-noise ratio (GSNR). GA-MEE reduces rate, but improves 10–15 dB statistics. For MSEMEE (GA-MSEMEE) algorithm, when GA-MSEMEE keep its improved dB, maintain approximately accuracy, nearly 100 iterations. addition, applied cancellation, average recovery proposed 7 points lower than other algorithms. validate effectiveness superiority environment, providing new methods deal multi-channel networks.