Rethinking biased estimation

In order to improve the accuracy of unbiased estimators used in many signal processing, there are methods that can be used. This approach is based on introducing a bias as a means of reducing the mean-squared error (MSE). In seeking unbiased estimators that perform well, it is typically accomplished by determining the minimum variance unbiased (MVU) estimator, using the theory of sufficient statistics of the Cramér-Rao lower bound. Estimators can be derived to outperform existing approaches for short data records and low signal-to-noise ratios (SNRs). The applications include the design of estimation algorithms for sonar, radar, and communications, as well as many other disciplines that rely heavily on precise measurement of parameters.