Advanced Traffic Modelling for Smart Cities - Project Demonstrator Using EVEREST Methodology

Precise traffic model helps municipalities to understand their everyday every minute traffic conditions, which is the result of its actual infrastructure capacity and mobility demand patterns of its citizens. Only with this knowledge, the municipality can take proper measures to improve its traffic conditions.The demonstrator is about a precise and fast calculation of a 3D traffic model from floating car data (FCD) with a use of traffic simulator. The demonstrator shows HPC computation in the data pre-processing step, where the collected FCD data are map-matched on a road network using accelerated calculation of Markov model and Dijkstra algorithms producing road speed profiles. In the next step, the deterministic traffic simulator consumes the speeds to calculate the 3D traffic model with the help of origin-destination matrix (ODM) input. The simulator makes use of demanding Probability Time-Dependent Routing (PTDR) algorithm to infer correct traffic delays. The demonstrator is being solved within EVEREST project. The project aims at developing a holistic design environment that simplifies the programmability of High Performance Big Data analytics for heterogeneous, distributed, scalable and secure systems. Everest proposes designs of computational heterogenous environment in the form of virtualised nodes with attached resources such as CPUs or FPGAs. Those nodes are orchestrated by HyperQueue scheduler that distributes tasks among them. Each node is further runtime optimized by the mARGOT autouner, which dynamically selects the most suitable version of a precompiled kernel based on the current state and resources available. Kernels are the application pieces compilable to various hardware architectures.

---

Contributors:

• Martin Surkovsky (IT4Innovations, VSB - Technical University of Ostrava)
• Stanislav Bohm (IT4Innovations, VSB - Technical University of Ostrava)
• Radim Cmar (Sygic)
• Jan Martinovic (IT4Innovations, VSB - Technical University of Ostrava)
• Gianluca Palermo (Politecnico di Milano)
• Katerina Slaninova (IT4Innovations, VSB - Technical University of Ostrava)