Position Finding with Prior Knowledge of Covariance Parameters

Butterly [1] presents a Bayesian approach as an alternative to the classical methods for solving the position-finding problem. Butterly assumes that bearing errors are independent and normally distributed with known variances. In the paper, the assumption of known variances is relaxed and it is shown that uncertainty about these variances can be incorporated into the model while also retaining the computational advantages of the Butterly formulation. It is also shown that the Bayes estimate and the classical maximum likelihood estimate will agree in certain cases.