Distributed Multisensor Fusion Algorithms for Tracking Applications
Abstract
The objective of the research under this ONR award is to develop multisensor fusion algorithms and sensor management techniques for tracking applications. Under this award, we have achieved a number of results: (1) We have developed a method of distributed fusion that is amenable to general distributed architectures; (2) We have developed non-simulation techniques for comparing multisensor fusion algorithms that are significantly more computationally efficient than performing Monte Carlo simulation evaluations; (3) We have investigated and compared the computational complexity and tracking performance of sequential and parallel implementations of multisensor fusion algorithms; (4) We have investigated the order of processing sensors of unequal qualities in sequential implementations of multisensor fusion algorithms; (5) We have developed several schemes for controlling sensor information and have evaluated the effects of sensor request delays; and (6) We have investigated the application of ordinal optimization and super-heuristic techniques for developing efficient implementations of our new sensor management methods. Our results have provided insight as to the relative performance of various multisensor fusion methods, and the results have also provided a basis for assessing the tradeoffs between performance and computational and communication requirements when planning new sensor network architectures or communication link protocols.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 2000
- Accession Number
- ADA377900
Entities
People
- Lucy Y. Pao
Organizations
- University of Colorado Boulder