An Event-Triggered Hybrid Consensus Filter for Distributed Sensor Network

Xiangxiang Dong,Giorgio Battistelli,Luigi Chisci,Yunze Cai
DOI: https://doi.org/10.1109/lsp.2022.3183494
2022-01-01
IEEE Signal Processing Letters
Abstract:An event-triggered consensus filter is proposed in this letter for state estimation in distributed sensor networks based on the hybrid consensus on measurement and consensus on information scheme. For bandwidth reduction and energy saving, an event-triggered transmission strategy is developed in which each node selectively transmits only the most relevant data so as to reduce data transmission while preserving the filtering performance. Two different transmission tests are performed in parallel, respectively on the prior and on the likelihood information pair, to evaluate the information loss (measured in terms of Kullback-Leibler divergence) that would be incurred if the current values were replaced by the predicted ones according to the last transmitted data. Simulation results on a distributed target tracking case-study demonstrate outperformance of the proposed filter with respect to conventional triggered filters.
What problem does this paper attempt to address?