First-Order Linearization of Saleh-Modeled Power Amplifiers For Improved Ofdm Performance in Rayleigh Fading Channels.

The main issue that affects multi-carrier communication systems occurs when power amplifiers produce non-linear distortion which deteriorates the signals that the receiver needs to detect. The Saleh model serves as the selected model for this research to represent PA nonlinearity. The first-order Taylor series linearization method applies to the RMS operating point to correct for the AM/AM and AM/PM distortions. The research testing proposes its approach in a 16-QAM OFDM system which uses MATLAB to simulate a Rayleigh fading channel with 20 dB SNR across 10 separate test runs. The Author’s used maximum stress conditions to assess the linearization performance by operating the PA at zero input back-off to achieve deep saturation. The Saleh model without compensation showed a baseline SER of 96.3% and EVM of 302.56% after the test. The EVM values from this measurement system exceed the 100% limit yet they match the mathematical requirements for RMS EVM when the average error power surpasses the power of ideal signals under extreme distortion conditions. The linearized system demonstrates significant enhancements which achieve statistical significance through improvement of SER by 3.84%, EVM by 17.71% and MER by 11.79% (p < 0.05). The t-distribution with 9 degrees of freedom provides the basis for calculating 95% confidence intervals which show mean improvements of SER: [2.5%, 5.2%], EVM: [15.1%, 20.3%], and MER: [10.1%, 13.5%]. Constellation diagrams provide visual evidence that the signal integrity has improved because the symbols cluster more tightly and tapping distortion has decreased after linearization. The linearization method retains operational capability across all standard FFT sizes because its implementation used an unconventional FFT size of 100 which diverged from the common power-of-two requirements. The method provides a digital predistortion solution that requires less computational power because it enables basic linearization operations at low system load. The first order system design restricts performance under severe nonlinearities and memory effects.