A New Adaptive Steepest Descent RLS Multi-user Algorithm

In order to solve problems as high error diffusion and instable convergence of Recursive Least Squares (RLS) algorithm in Code Division Multiple Access (CDMA), this paper combined local steepest drop characteristic of steepest descent algorithm presented a stable convergence and lower complexity algorithm -- D-RLS in condition of high multiple access interference (MAI) and intersymbol interference (ISI). New algorithm effectively reduce the diffusion of detection errors and ensure the local convergence in detection process, therefore achieve rapid global convergence to a desired user and improved dynamic tracking capability. The simulations shown that, new algorithm has better convergence property and dynamic stability than original RLS algorithm.