Toolset for Content-Based Automated Communication Analysis

Teams play an increasingly critical role in complex military operations in which technological and information demands require a multi-operator environment. However, verbal communication, which represents a rich source of data on team cognition and collaborative thought, is notoriously difficult to analyze. We propose to develop a set of tools for the automatic analysis of verbal communication data in order to assess and monitor individual and team knowledge. The toolset uses a Latent Semantic Analysis (LSA)-based intelligent software agent for embedding automatic, continuous, and cumulative analysis of oral interactions in individual and team operational environments, It further includes additional statistical measures of sequential team behavior and a running task model of team cognition and process behavior (C3TRACE). In Phase I, the project will identify existing communication analysis techniques, test these techniques on existing communication data and demonstrate and report on their reliability, validity and effectiveness at quantifying aspects of team performance and situation awareness. Phase II and beyond will develop the techniques into a robust system for automatically monitoring and providing feedback on teams and individual performance and test them within operational contexts. The full system provides a complete communication analysis pipeline, automatically converting speech into performance measures in real-time.