Delay/Doppler estimation in a multipath environment
Typical Naval test and evaluation exercises involve the transmission of known acoustic signals by the vessel under test. In the past, most systems were designed for water depths in excess of 2000 feet. In this case, the acoustic signal travels via a direct path to hydrophones mounted on or near the ocean bottom. Since the end of the Cold War, the Navy has shifted its efforts to regional conflicts usually requiring the use of vessels in littoral waters. In this case, the acoustic signals do not necessarily travel directly to the receiver. Instead, they may reflect from the surface and/or bottom several times before reaching the receiver. These reflections distort the signal by introducing phase shifts and constructive and destructive interference due to overlaps caused by the difference in arrival times.^ Most of the current delay estimation methods are iterative and tend to become ill-conditioned, or perform poorly in the presence of Doppler. This thesis provides a background in the currently accepted underwater vehicle tracking processor which estimates time delays of a known Differential Phase Shift Keyed (DPSK) Signal. It is shown that this processor performs well in high Signal-to-Noise Ratio (SNR), low Doppler geometries; its performance deteriorates considerably elsewhere.^ An algorithm for computing the Cramer-Rao Lower Bound (CRLB) for any number of specular multipaths in a discrete-time signal is provided. This algorithm can be implemented in either the time domain (exact signal model is required) or frequency domain (only Doppler scaled replicas of the signal are required). The CRLB is computed for the examples in future sections.^ The Manickam/Vaccaro noniterative least squares time delay estimation method is extended to the delay/Doppler case under which the original algorithm is shown to provide erroneous results. This algorithm requires the selection of a two-dimensional indicator set which encompasses the true paths. Examples of simulated three path conditions are shown to provide better estimates than the delay only case. However, the indicator set is shown to be too large.^ An algorithm for selecting the appropriate indicator set for two closely spaced, equal amplitude paths is provided. It is shown to have significant improvement over the rectangular set and approaches the CRLB. ^
Engineering, Electronics and Electrical|Engineering, Marine and Ocean
Amir Wadi Habboosh,
"Delay/Doppler estimation in a multipath environment"
Dissertations and Master's Theses (Campus Access).