Enhanced digital predistorter based on normalized least mean square and particle swarm optimization algorithms

In this paper, two modified digital predistorters based on particle swarm optimization and normalized least mean square algorithms are introduced. The proposed algorithms reduce the non-linearity and memory effects that cause in-band distortions and increase the spectral regrowth in adjacent channels. the first proposed algorithm solves the convergence guarantees by the local search caused by particle swarm optimization. The second algorithm solves the same problem and improves the full random distributed particles defined in the swarm, simultaneously. Simulation results of the two proposed algorithms are demonstrated and compared with the conventional particle swarm optimization algorithm, traditional normalized least mean square algorithm and also compared with each other.