This project will explore the potential of neural network and wavelet algorithms to yield significant performance improvements needed for future acoustic detection and classification systems. Future systems must automatically recognize complex acoustic signatures in realistic environments. Neural and wavelet algorithms open new possibilities for designing sonar systems by offering the potential for compact, massively parallel computational hardware with increased flexibility in accommodating new targets and environments. Computer simulations will be employed to investigate neural and wavelet processing algorithms that are expected to yield significant increases in the accuracy with which specific acoustic signals can be detected, classified and tracked.
