Multiple Time Series Node Synchronization Utilizing Ambient Reference
Abstract
Special Operators, like professional athletes, must maintain peak physical performance in and out of training, and like professional athletes, they could benefit from having real time biometric data used to improve training and performance. In professional sports, the use of wearable sensor networks may be soon became commonplace in gathering training data for performance evaluation, but the same is not truefor Special Operators, because of their specialized and exclusive training, and the demanding operational conditions. One major requirement for meaningful data analysis and collective signal processing targeted to performance assessment, is the need for fine scale synchronization among communicating nodes and across multiple domains. The severe requirements that Special Operators have to obey do not allow for common synchronization techniques, such as those that rely on a global clock. For this reason, techniques that rely on a different synchronization mechanism could potentially be implemented into a practical solution to provide real-time feedback data during training orin field operations. This final report summarizes: 1) the development of a set of Naive Bayes classification algorithms with their application to a validated (PAMAP2 dataset), 2) development of a generic analog sensor node capable of acquiring physiological data for analysis, and 3) implementation of a Data Explorer and Labeling GUI to enable comparison and temporal alignment of extracted features of interest.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 31, 2014
- Accession Number
- AD1001046
Entities
People
- Alessio Medda
- Shean E. Phelps
Organizations
- Georgia Tech Applied Research Corporation