Unified Collection and Coordination for UCAVs
Abstract
Lockheed Martin Tactical Systems (LMTS) and Scientific Systems Company, Inc., (SSCI) are conducting a research program aimed at the problem of managing autonomous, self-reconfiguring swarms of intelligent sensors and weapons. Their goal is a theoretically comprehensive but also potentially practical unification of the two major aspects of the problem: Multi-Agent Collection (i.e., distributed robust data collection, fusion, and interpretation of often poorly characterized or poorly defined data) and Multi-Agent Coordination (i.e., distributed robust platform/sensor monitoring and control). Data fusion/correlation and sensor/platform management is a control theory problem in which the underlying entities (targets, sensors, data, platforms) are stochastically varying multi-object systems. This necessitates a practical unification of control theory and point process theory. LMTS developed a potentially tractable approach to multisensor, multitarget sensor management. This approach uses approximate multitarget filters (a probability hypothesis density (PHD) filter or a multiple-hypothesis correlator (MHC) filter) as the underlying multitarget filter/predictor, and "natural" probabilistic objective functions (e.g., posterior expected number of targets) combined with a "maxi-PIMS" optimization strategy for tractably hedging against unknown future observations. The approach was extended to multi-step, look-ahead sensors with non-ideal dynamics (e.g., UAVs), sensors without directly observed states, and communications drop-outs. A list of the author's technical publications, conference proceedings and presentations, technology transitions, and honors is included.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 28, 2004
- Accession Number
- ADA425515
Entities
People
- Ronald P. Mahler
Organizations
- Lockheed Martin