Managing data incest in a distributed sensor network

Multisensor data fusion is one of the key evolving technologies that can be implemented on several architectures. Distributed sensor network architecture provides a good balance between cost, scalability and communication limits of links connecting multiple sensors. However, this architecture is prone to the problem of data incest. Data incest arises due to multiple usage of identical information as if it were independent information. In most cases, it will falsely increase the confidence of a biased overall estimate. It is a fundamental issue in the many-to-many sensor network configurations, where all network nodes are communicating with all nodes. This paper describes a fusion strategy which can be adopted for this distributed network structure that renders incest free estimates.