Maximum likelihood methods for difficult tracking problems

This paper discusses the advantages, disadvantages and methodology of maximum likelihood estimators (MLEs) applied to tracking problems. The paper goes on to explain how a criterion derived by Akaike can, in conjunction with the maximum likelihood fit, be used to help optimise the size of the vector of unknown parameters representing the target kinematics. Some of the concepts discussed are illustrated with numerical results relating to a simple bearings-only passive tracking problem.