Understanding the future transportation infrastructure performance demands a smart Cyber-Physical Systems (CPS) approach integrating heterogeneous sensors, versatile computing systems, and mobile agents. However, due to sensor versatility and computing intricacy, designing such systems faces challenges of immense complexity in mobile sensor fusion, big data handling, system scalability, and integration. This paper introduces SIROM3, a Scalable Intelligent ROaming Multi-Modal Multi-Sensor framework, for next generation transportation infrastructure
performance inspection. SIROM3 offers a scalable and expandable framework through orthogonally abstracting software / hardware structures in a layered Run-Time Environment (RTE), which facilities sensor fusion, distributed computing, communication and mobile services. A Heterogeneous Stream File-system Overlay (HSFO) and a flexible plugin system (PLEX) are embedded in SIROM3 to simplify big data storage, processing, and correlation. To evaluate the scalability of SIROM3, we implemented a mobile sensing system of 30 heterogeneous sensors and 5 computing platforms coordinated by 1 data center. SIROM3’s expandability is highlighted by adding an advanced radar platform
which required less than 50 lines of C++ code for integration. Over 20 terabytes of data covering 300 miles have been collected, aggregated, and fused using SIROM3 for comprehending the pavement dynamics of the entire city of Brockton, MA. SIROM3 offers a unified solution and ideal research platform for rapid, intelligent and comprehensive
evaluation of tomorrow’s transportation infrastructure performance using heterogeneous systems.

Zhang_Compsac2014.pdf