Passive direction of arrival (DOA) tracking

Abstract

The work reported in this thesis is concerned with the Passive Direction Of Arrival (DOA) tracking problem. Traditionally model based approach is used to solve tracking problems. The measurements (DOA) are nonlinearly related to position (element of state model) and does not include any range information. This makes the tracking problem very complex. Since we also have to deal with the effects of noise and sensor uncertainties, this results in a potentially unobservable nonlinear estimation problem.

Estimation of position of target from DOA measurements which has nonlinear relation with state vector, demands the use of a linearized model approach.

The whole thesis work has been divided into two part. First part analyzes the observability problems of DOA tracking i.e. it analyze the given scenario for the unique solution and suggest some correction to get proper system parameter estimation. This part also tries to find the principle causes of filter divergence.

In second part, the thesis is concentrated on the estimation methods used to track target. In this part thesis analyzes two different estimation algorithms named Extended Kalman Filter(EKF) and Pseudo Linear Estimator(PLE).

Thesis proposed a Triangulation Estimator which estimates the target motion parameter by using the geometry formed by the target and sensors available. Finally thesis compared the performances of the algorithms considered in it on the realistic underwater scenario.