A novel approach to pruning the general Volterra series for modeling power amplifiers

This paper proposes a recursive optimum-term selecting (ROS) approach to pruning the general Volterra series, by which we can achieve a custom-tailored model to characterize nonlinearity of wideband power amplifiers (PAs) with memory effects. The achieved model is more suitable for the individual PA than those static models, such as the MP and GMP models, as it selects the most efficient terms from the general Volterra series based on the theory of recursive correlation cancellation. Simulation results show that the approach is effective, and the pruned model developed by the proposed approach is efficient as well as adaptable.