Shallow-Water Sensor Placement
Abstract
Placing acoustic sensors allows for remote detection of vessels in areas of interest. We study the problem of placing only a few sensors to effectively monitor signals at many locations. Though our approach applies to sensor placement problems generally, we focus on placing hydrophones to efficiently monitor acoustic signals over a large region. Our starting point is the mutual information criterion for sensor placement, which despite being theoretically attractive has not been widely adopted because of its computational difficulty. To remedy these computational challenges, we introduce a novel branch and bound algorithm that relies upon a new semidefinite programming relaxation of the mutual information problem. Our contributions allow practitioners to solve sensor placement problems to global optimality while exploring only a fraction of the solution space required by brute force. Our work has been made open source at https://github.com/rbassett3/mutual_info_sensor_placement.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2022
- Accession Number
- AD1201791
Entities
People
- Erik V. Vargas
Organizations
- Naval Postgraduate School