Behavioural modelling of RF nonlinear power amplifiers in orthogonal frequency division multiplex transmitter systems

The major weakness of frequently used Orthogonal Frequency Division Multiplexing (OFDM) signals is their high sensitivity to nonlinear amplification process when they are driven through RF power amplifiers (PAs). In research stuidies of this nonlinear distortion, PA behavioural models, as compared with physics models,are attractive for their low computation cost, good accuracy, and low application complexity in simulation. This thesis reports mainly on research studies of PA nonlinear distortion of OFDM signals. The topics treated include PA modelling techniques, PA and OFDM simulation algorithms, analysis of nonlinear PA distortion of OFDM signals, and PA linearization and linearisation benefits. The novel contributions to the field emanating from this research may be listed as follows: 1. Clarifications of modelling mechanism of classical Bessel Fourier (BF) model by using instantaneous voltage transfer (IVT). 2. A new Modified Saleh (MS) model for nonlinear PA behavioural modelling. 3. A new Modified BF (MBF) model for nonlinear PA behavioural modelling. 4. A new SAO (Statistics Applied on Output) simulation algorithm for the analysis of OFDM output signals after amplification by a nonlinear PA; based on the BF behavioural model. 5. Novel analysis of internal characteristics and properties of the intermodulation products caused by the nonlinear distortion of OFDM signals. 6. Novel insights to equivalency attributes of EVM and ACPR figures of merit (FOMs) measures in OFDM signal systems. 7. Contributions to the theoretical foundations of the percentage linearization measure and soft upper bounds (SUB) to linearisation benefits. 8. Validation of linearization benefits of WiFi signals through PAs with various percentage linearization (PL) degrees applied. The MBF is shown to be the best memoryless behavioural model under most headings including especially being the most accurate low coefficient count model. The MS model is shown to be 2nd best to MBF, and considerably better than the Saleh model both in modeling accuracy and range of PA characteristics it can model. The SAO algorithm technique is a new effective addition to the analysis tool when studying nonlinear amplification of OFDM signals. Its help when studying signal FOMs and upper bounds of linearisation benefits are shown.