On a suboptimal nonlinear filter

This paper examines a new suboptimal method for the recursive estimation of the state of a nonlinear differential system utilizing observations which are a nonlinear function of the states. Model smoothness, such as required with most existing filtering techniques, is not required here. The technique is shown to be promising for real-time target tracking of a moving rigid body by sonar in a noisy underwater environment as demonstrated by a numerical example.