Multi-sensor autonomous tracking for Maritime Surveillance

A novel approach to perform autonomous data correlation is described, focussing on the alignment of tracks originated by different sensors and technologies into a common spatial and time reference system, and introducing the quality characterisation of the association process based on Bayesian inference.

Abstract

The fusion of multiple monitoring systems is essential to build an accurate recognised traffic picture in support to maritime surveillance. In this paper, a novel approach to perform autonomous data correlation is described, focussing on the alignment of tracks originated by different sensors and technologies into a common spatial and time reference system, and introducing the quality characterisation of the association process based on Bayesian inference. This is performed following the sensors uncertainties characterisation, the implementation of the tracks propagation concept and the influence of the scenario of interest calculation. The proposed algorithm represents a flexible solution to the problem of data fusion on multiple platforms, either ground-, air- and space-based.