Nonlinear filtering lower bound evaluation of passive tracking systems

A new approach to the performance evaluation of a passive acoustic array tracking system is presented. The approach uses nonlinear filtering lower bound algorithms to estimate the RMS localization (range and bearing) errors and their sensitivity to signal-to-noise ratio (SNR), dynamic deformation of the array, and acoustic environment variations. The algorithms used model the stochastic nature of the problem and are effective at low as well as high SNR. The significant differences between the approach followed here and classical parameter estimation lower bounds (such as the Cramer-Rao bound) are discussed. A numerical example illustrating the methodology is presented.