Signal-dependent error in specular multipath estimation

The signal-dependent error for least squares specular multipath estimators is analyzed for active sensors. The error analysis also applies to other deconvolution techniques for multipath estimation such as projection onto convex sets (POCS) that uses the square error as a constraint. The results explain how the selection of the transmit signal affects the estimator performance. Example mean square error plots are provided for a linear frequency modulation (LFM) signal to demonstrate the results. Given the set of possible locations for multipath components, the optimum signal to minimize the square error can be selected using these results.