Coastal surveillance in multi sensor environment: A design approach

This paper discusses a design methodology for developing simple yet robust coastal surveillance application for tracking maritime objects in multi sensor environment. Hierarchical and distributed data processing architecture has been discussed. A linear Kalman estimator is used for target state estimation. A novel data association approach based on divide and conquer paradigm has been proposed. Track to track fusion, based on Bayesian Inference has been discussed. The paper also discusses approach to detect anomalies. A novel approach to manage inconsistencies, resulting from erroneous association and fusion of data which is unavoidable particularly in those scenario where maritime objects move in close clusters, has been proposed.