Volterra series based modeling and compensation of nonlinearities in high power amplifiers

Demands for higher data rates coupled with increased competition for the available RF bandwidth demand communications systems with greater bandwidth efficiency. Bandwidth efficient modulation techniques, such as QAM, require highly linear amplifier performance to achieve acceptable bit error rates. The nonlinear distortion which results when an amplifier is operated near saturation may preclude their use. One attractive option is to predistort the signal by placing a nonlinear filter in the signal path which compensates for the distortion introduced by the amplifier. In this paper, we present a predistortion technique which uses an inverse amplifier model based on a Volterra series approach. The input-output data from the amplifier is used to develop the parameters for a discrete Volterra series. The RLS adaptive filter technique is utilized to provide periodic updates to the inverse filter which allows tracking of amplifier variations. Results are presented for the case of 64 QAM modulation using a TWT amplifier operated at saturation