Observation Capability for Distributed Multi-Sensor Information Fusion

Abstract

In distributed multi-sensor system, many information fusion rules have been investigated to process the information data. And, it is a primary issue to study better fusion rules. In this paper, a statistical viewpoint that multi-sensor information fusion rules based on Fisher information are investigated. And we use them to analyse the observation capability. Firstly, a distributed multi-sensor observation model from the viewpoint of Fisher information is established. Then, three fusion rules represented by the amount of information, i.e., independent observation, information fusion and signal fusion, are proposed. In particular, we prove their ralations in detail and apply them to fuse multisensor information to obtain different observation range. Finally, the observation range obtained by signal fusion and information fusion are compared with that obtained by conventional coherent integration and non-coherent integration. Simulation results show their consistency.