Formal convergence guarantees for DGKCF

Establish formal convergence guarantees for the Dynamic Generalized Kalman Consensus Filter in sensor networks with limited sensing ranges and switching communication topologies.

Background

The paper introduces the Dynamic Generalized Kalman Consensus Filter (DGKCF), which replaces the fixed consensus step size of conventional weighted consensus schemes with locally computed information-based weights derived from neighboring agents’ prior error covariance matrices. The reported evaluation is simulation-based and demonstrates improved RMSE, MAE, and convergence time relative to KCF and GKCF under switching topologies.

A formal proof that the proposed algorithm converges is not provided. Establishing such guarantees would theoretically validate the stability and asymptotic behavior of DGKCF under the switching communication topologies and heterogeneous observability conditions considered by the framework.