Implementation of a Maximum Likelihood Doppler Spread Estimator on a Model-Based Design Platform

Abstract

A new maximum likelihood (ML) Doppler spread estimator, recently shown to outperform most representative state-of-the-art solutions both in accuracy and complexity, is implemented on a FPGA-based platform. Rapid prototyping of the entire design is built as a highlevel Simulink model using Xilinx System Generator IP blocks. The RF front-end of the design is implemented using the Nutaq's model-based design kit (MBDK). The ML Doppler spread estimator is assessed at a sampling rate of 80 Msps over a realistic RF channel generated by the EB Propsim FS8 channel emulator. Comparisons with the original floating-point MATLAB version suggest negligible performance losses, thereby validating and confirming the efficiency of the new real-time overt-the-air hardware design and implementation.