Cosense: Collaborative Sensemaking of Distributed Data for Recognition and Condition Monitoring
Abstract
This report summarizes PARC's work on collaborative sense making under the DARPA SensIT program. The report is divided into two parts. Part I summarizes the original contract work which was centered on topics such as distributed multi-level feature analysis, distributed hypothesis management, and scaling metrics for distributed sensors. An extensive bibliography of the detailed papers resulting from this work is provided. Part II describes the work under the two contract extension where we proposed a distributed attention approach, spreading tasks through the network and filtering out the subset of useful data close to the data sources, proceeding from signals to symbols as we progress from data source to end user. Our architecture separates concerns of how and where to acquire the data from issues of processing. Data interpretation is separated into layers with clear interfaces. Testbed and simulations of the architecture and various algorithms are documented here.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2005
- Accession Number
- ADA435466
Entities
People
- James Reich
- Juan Liu
- Maurice Chu
- Peter Cheung
Organizations
- Xerox