No abstract

No abstract

We present a general method for approximating the global error performance of ML type multi-parameter estimators. The method is a multi-dimensional extension of the approaches of 121 and [l]. Compariaons between the global approximation to the MSE for known and unknown "nuisance" parameters indicates the degree to which nuisance parameters degrade the performance of ML estimators. The results of a numerical study illustrate the implementation of our method for time dday estimation in cases of coherent interference and doppler.

No abstract