| Popis: |
Most signal processing algorithms are based on the frequency‐domain cross‐spectral density matrix, which is constructed from a large time window of data from an array of hydrophones. Forming this matrix involves averaging the data, a process that wastes a great deal of potentially useful information. By working with an optimal time‐domain beamformer, which uses all of the information in the received time series, it is possible to determine the bearings and signatures of plane‐wave sources with a small time window of data and relatively few hydrophones. Due to the recent development of efficient methods such as simulated annealing for solving optimization problems involving many unknown parameters, optimal time‐domain beamforming has the potential to become a very useful approach. Simulated annealing is analogous to slowly cooling a pure liquid to obtain a perfect crystal. For the optimal beamformer, the unknown bearings and time series appear to be analogous to a mixture of liquids of different freezing p... |
