Adaptive Models and Fusion Algorithms for Information Exploitation

摘要

The main aims for the project were to develop methodologies for managing and exploiting information available from multiple heterogeneous sensors/sources under limited sensing, computation and communication capabilities. Towards these goals, we conducted research along four directions, viz., source querying strategies, information fusion algorithms, learning algorithms to model the changing nature of data sources, and algorithms to exploit spatiotemporal relationships between different sources. We addressed realistic scenarios, with constraints on communication and computational resources, and characterized by time-varying and unpredictable changes in environments with spatially mobile entities. In many such problem scenarios, the information gathering and analysis efforts are complicated by the fact that data sources may be faulty and unreliable. This motivated addressing the tasks of situation assessment using asynchronous, heterogeneous and uncertain data sources. Results obtained have been documented in a number of technical publications.