This paper presents a theoretical approach that has been developed to capture the computational intensity and computing resource requirements of intelligent transportation system (ITS) data and analysis methods. These requirements can be transformed into a common framework, region-based divisions of ITS computational data processing, which supports the efficient use of cyber infrastructure. The computational transformation is performed to characterize the computational intensity of a particular ITS data analyses. The application of the theoretical approach is illustrated using two ITS data processing methods: multi-sensor data fusion by integrating federated Kalman filter and D-S evidence theory, and geospatial computation on GPS data for urban traffic monitoring. Through the application, the development of region-based division method is decoupled from specific high performance computer architecture and implementations, which makes the design of generic parallel processing solutions feasible for ITS data analyses. The experimental results show that the framework can be applied to divide the ITS data analyses based on regions into a balanced set of computing tasks, and parallelizing data fusion and geospatial computation algorithms achieves better speedup.
展开▼
