摘要: |
Tracking the movement of individuals in complex urban environments using mobile sensors is a challenging, but important problem in applications such as law enforcement, homeland security and defense. The Dynamic Data Driven Application Systems (DDDAS) paradigm offers a natural approach to attacking this problem. This two-year research project explored new computational technologies based on the DDDAS paradigm that could be applied to track vehicles in real time. Research accomplishments from this project include (1) the development of approaches to improve the transient response of data driven distributed simulations, (2) development of methods for on-line data driven calibration of traffic simulations, (3) development of data analytics for real-time prediction of vehicle trajectories, (4) development of algorithms for efficient execution of replicated transportation simulations, (5) analyses of data distribution methods, (6) energy analysis of synchronization algorithms for distributed simulations, and (7) development of parallel algorithms for non-negative matrix factorization for vehicle detection. |