Analysis and Algorithms for Imperfect Sensor Deployment and Operations
Abstract
The research begins by considering a class of problems in which a set of sensors has been deployed across some theater to monitor a set of targets, where sensors are subject to uncertain operation. In particular, it is usually appropriate to assume that a limited number of sensors can fail and erroneously report targets, or fail to detect targets that are indeed present. Simply using a redundant covering of targets is unnecessarily expensive, and thus more sophisticated methods are required to robustly cover a network in the presence of faulty sensors in a minimum-cost manner. This gives rise to the problem of minimizing the maximum number of "ambiguous nodes," i.e., nodes at which it is impossible to determine whether or not a target exists given the set of sensor readings. This core problem requires the invention of new network interdiction strategies, because the innermost recourse optimization problems in our scheme are integer programs, rather than linear programs (which can be dualized and solved as one monolithic integer programming problem). Therefore, this project explores applications and general theory that have clear benefit to the Air Force. This funding has created vital research opportunities for graduate students in the investigator's departments.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 23, 2016
- Accession Number
- AD1011571
Entities
People
- J. Cole Smith
- Joseph Guenes
Organizations
- University of Florida