CHAMPS: a near-ML joint Doppler frequency/TOA search for channel characterization

The paper introduces a frequency-domain near-maximum likelihood (ML) joint Doppler frequency (DF)/time of arrival (TOA) search for channel characterization. The search is iterative and operates by finding one or more DF/TOA pairs at each step that minimize the channel estimation mean squared error. Because of its excellent ability to estimate frequency-domain channels as well as identify channel parameters (i.e., DFs and TOAs), we call the algorithm CHAMPS (CHannel estimation via an Approximate ML Parameter Search). CHAMPS is used to characterize the channels measured from an experimental 18.775 MHz OFDM mobile system operating at 3.676 GHz. The experimental system was used to collect data in Schaumburg, Illinois, an environment characterized by both industrial and residential areas. Differential delays up to 24 μsec were detected as well as Doppler frequencies that are three times the expected Doppler frequency due to reflections off vehicles traveling in the opposite direction as the test van.