Reproduce “Distributed Maximum Likelihood for Simultaneous Self-localization and Tracking in Sensor Networks” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Nikolas Kantas, Sumeetpal S. Singh, Arnaud Doucet

The authors propose decentralized online maximum-likelihood methods that use observations of a shared moving target to estimate sensors’ relative offsets in local coordinate frames without GPS or direct neighbor-distance measurements. Nodes exchange compact messages. In an 11-node tree simulation with a linear-Gaussian model, plotted errors converged in around 1,000 RML iterations and approximately 2,000 online EM iterations; this is empirical, not a general guarantee.

Key insights

  • Simultaneous observations of a common moving target by multiple sensors provide sufficient correlated information to infer their relative positions without external localization infrastructure.
  • Robust simultaneous localization and tracking can be achieved in a decentralized and on-line manner by integrating recursive ML updates with a distributed Kalman filter or Extended Kalman Filter.
  • A scalable message passing mechanism can effectively aggregate global network information at each node, making distributed ML estimation computationally efficient and suitable for large networks.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.