Comprehensive Tactical Network State Inference from Incomplete Data
Abstract
The objective of this proposal is to utilize recent developments in statistical learning to derive cognitive radio network state acquisition, tracking, and prediction, even under dynamic battlefield operation. The vision is to construct a comprehensive cognition infrastructure, capable to cope with dynamic operating conditions of tactical ad hoc wireless networks, as well as incomplete, corrupt, and sporadic data, which reflect the cost and restrictions in acquiring relevant state measurements. The proposed research is divided into three interrelated thrusts: 1. Comprehensive tactical network state inference at the PHY layer; 2. Comprehensive state inference at the MAC and networking layers; 3. Anomalography for intrusion-resilient tactical network operation. Research thrust area 1 puts forth nonparametric basis pursuit (NBP) as the overarching tool embracing a number of kernel-based learning and inference techniques, which will be adapted to acquire and maintain network states at the PHY layer. Dynamic NBP extensions are explored in thrust 2, where the state of the network comprises band occupancy to record whether certain frequencies are under use by adversarial units, queue lengths, traffic loads, and communication delays among tactical units. Thrust 3 is geared towards threat protection by invoking alarms upon the occurrence of intrusions or attacks.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 12, 2017
- Source ID
- W911NF1510492
Entities
People
- Georgios B. Giannakis
Organizations
- Army Contracting Command
- United States Army
- University of Minnesota