Dynamic Information Collection and Fusion
Abstract
The project is aimed at developing a comprehensive framework for control of information collection, fusion, and inference from diverse modalities Our research has been organized under three inter-related thrusts. The first thrust addresses system modeling and local information processing. The second thrust emphasizes the interaction between information and control at different abstraction levels. The third thrust is focused on decentralized processing and interactive fusion. Within Thrust 1. we focused on exploring, developing, and utilizing mathematical models for hard and soft observations, the physical and abstract information states, and the sensing state for local information processing and inference. Within Thrust 2. we have obtained a number of results on sensor scheduling for target tracking, controlled sequential multi hypothesis testing, controlled sensing for graph classification, universal outlier hypothesis testing, and optimal search and stop. Within Thrust 3, we have obtained a number of results on interactive fusion, data reduction with quantization constraints, network consensus and quantized alternate direction method of mutlipliers (ADMM), and resource management in sensor networks.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 02, 2015
- Accession Number
- AD1003771
Entities
People
- Biao Chen
- Prakash Ishwar
- Venugopal V. Veeravalli
Organizations
- University of Illinois Urbana–Champaign