Research areas

  • Target tracking and multitarget estimation
  • Switching-state estimation and hybrid systems (IMM, GPB)
  • Data association under measurement-origin uncertainty (PDAF, JPDAF, MHT)
  • Sensor fusion for radar and sonar systems

Key papers

Recent work

(To be filled by future ingests.)

Collaborators

  • E. Mazor
  • A. Averbuch
  • J. Dayan
  • H. A. P. Blom (Blom & Bar-Shalom 1988, original IMM paper)

My notes

Founding figure of modern target tracking. The Bar-Shalom textbooks (“Estimation with Applications to Tracking and Navigation”, 2001; “Tracking and Data Association”, 1988) are the canonical references for the field, and the IMM algorithm — co-developed with Blom in 1988 — became the de facto standard hybrid-state estimator for maneuvering-target tracking by the late 1990s. For the present asset-pricing project, the relevance is the IMM/RBPF contrast: both exploit the same Rao-Blackwellization (continuous state is linear-Gaussian conditional on the regime path), but IMM uses deterministic moment-matching at every step while RBPF uses stochastic sampling-with- resampling in regime-history space. The Bar-Shalom corpus is the right starting point for understanding the cost/accuracy tradeoffs of every switching-state filter ever proposed.