Bayesian sensor registration for multistatic active sonar

Multistatic tracking presents unique challenges not found in more traditional monostatic or bistatic settings. Important among these is the problem of sensor registration. This paper examines the problem of sensor registration for the case of a distributed system of active sonar sensors. The problem is formulated in terms of Bayesian measurement likelihood functions with particular emphasis on cases in which measurements give only partial information about the state variables. Estimation for two particular sets of state variables is considered: (1) source position and velocity and (2) time and compass offsets.