Partner: Sygic, a.s.

Field: information technology, Smart City

The Smart City is a concept aiming to increase the quality of many aspects of our life in modern cities with respect to reduced energy consumption. These include, for example, traffic optimization, efficient control of air-conditioning and heating in buildings, easier access
to public services, etc. In the field of traffic management, for instance, cities need to know what would happen
if a road was closed for bridge reconstruction work, the road narrows, or if there is overload of the road network
because of a mass sports event. Such information is valuable for authorities responsible for transport infrastructure operation in managing potential crisis
situations, and in efficient planning of road network repairs at an appropriate time so as to minimize their impact on traffic flow.

In cooperation with the navigation company Sygic, a.s., we are focused on the aspect of traffic optimization
in cities in the form of developing specific services based on advanced data analyses and simulations as well
as an expert, self-adaptive navigation system combining client and server-side routing. This system uses historical data, up-to-date traffic flow monitoring,
and graph theory based methods for determining the potentially busiest sections within the transport network.
Furthermore, the system is extended with integration of population mobility data. For macromodeling within the
system, we use our own implementation of a betweenness centrality algorithm, which is being further improved through conducting research and development. Application of this algorithm
to large graphs, which represent a large number of road segments, is very computationally intensive and therefore we use the supercomputing infrastructure so as to provide up-todate information to users as fast as
possible.

Efficient provision of such created services is secured using the technologies developed within the ANTAREX
project under the EU H2020 program. The main objective of ANTAREX is to provide methods for mapping, real-time administration, and autotuning for HPC systems with respect to efficient energy consumption. In order to do so, tools based on LARA, a specific domain-oriented language for heterogeneous systems, have been developed. In combination with autotuning, it is thus possible to efficiently apply the principles of parallelization, scalability, and dynamic self-adaptability to fully utilize the power of an HPC system. The results generated within this cooperation have been applied within the ANTAREX project, published in a number of articles, and last but not least presented to the general public at conferences including SC 16 (the International Conference for High Performance Computing, Networking, Storage and Analysis) held in Salt Lake City as well as at Transportation Research
Arena 2018 in Vienna.

http://www.antarex-project.eu

PARTNER´S NOTE

Radim Cmar
Sygic, a.s.

Our cooperation with IT4Innovations and taking advantage of their
experience and resources in the field of supercomputing and traffic
modeling extend our skills and provide large-scale transport analysis for cities in near real time, which makes our product very competitive on the current global market.