作者: Reza Olfati-Saber , Jeff S. Shamma , Emilio Frazzoli , Elisa Franco
DOI:
关键词:
摘要: In this paper, we address distributed hypothesis testing (DHT) in sensor networks and Bayesian using the average-consensus algorithm of Olfati-Saber & Murray. As a byproduct, obtain novel belief propagation called Belief Consensus. This works for connected with loops arbitrary degree sequence. consensus allows computation products n beliefs (or conditional probabilities) that belong to different nodes network. capability enables broad variety applications. We show admits Lyapunov function quantifies collective disbelief benefits from scalability, robustness link failures, convergence under variable topology, asynchronous features algorithm. Some connections between small-word speed are discussed. A detailed example is provided detection multi-target formations The entire network capable reaching common set associated correctness hypotheses. demonstrate our DHT successfully identifies test formation sensors self-constructed statistical models.