Feedforward Power Amplifier Adaptive Control System Design And Fpga Implementation

With the rapid development of modern mobile communication, high channel capacity and wide frequency bandwidth are required to be provided by communication systems. However, spectrum resources are becoming increasingly scarce for the sharp increase of subscribers, and the only method is to improve spectrum efficiency through linear modulation technique. Due to the wide use of linear modulation technique, the linearity of power amplifiers have become increasingly demanded for mobile communication systems. To maintain adequate channel capacity and avoid adjacent channel interference, external linearization circuitry must be added to correct the distortion generated by the amplifier. The commonly used linearization techniques for power amplifier include power back-off, negative feedback, pre-distortion and feedforward etc. The only available method providing both wideband (multiple MHz) and high linearity is feedforward. However, because of its high sensitivity to temperature, aging, different frequency channels and other variations, feedforward requires automatic adaptation of key parameters for reliable distortion cancellation as operational and environmental conditions vary. So this linearization technique is expensive, complex, and very difficult to implement.Firstly, the principle of nonlinear distortion produced by power amplifier is presented. Then we analyze several commonly used linearization techniques for amplifier, and compare their advantages and disadvantages respectively. The thesis focuses on the control theory of adaptive feedforward control system and the algorithm for adaptive control parameters. We have taken the least signal correlation method as the optimization criterion in our adaptive control system to overcome some difficulties, such as incorrect convergence caused by DC offsets at the output of analog mixers and masking of weak signals by stronger ones. In the design of down-conversion module, we improve the commonly used digital down-conversion method with reference to the U.S. Patent 5157345, and obtain the narrow band audio signals, thereby greatly reducing the performance requirements of the system hardware.We have simulated the adaptive feedforward control system with two-tone test signal in MATLAB. The result verifies that the variable loop step LMS algorithm can effectively control the parameters'convergence from their initial values to their optimal solutions respectively, and depicts the process of distortion cancellation. Finally, a board-level experimental circuit based on FPGA is designed and tested for our adaptive control system, and then the logical functional simulation and implementation of LMS algorithm in FPGA are completed. The final simulation results reflect the feasibility of our design.