5G Driving Trainer (5G-DT)


Elios srl,

Experiment description

The 5G Driving Trainer (5G-DT) system consists of a proof of concept prototype able to perform real-time image processing of moving vehicles and elaborating solutions for safe, correct and ecological driving behavior.
The experimentation includes VNF instances for the 5G-DT data collection and processing along with the 5G-DT edge computing system, deployed at the edge of the infrastructure. Such an hybrid architecture permits, on the one hand, to efficiently process video streams closer to the sources and on the other hand, to take advantage of the distributed cloud 5GINFIRE infrastructure for global data aggregation and data fusion.
Interaction with OBU is experimented to evaluate the overall system performance and capabilities including the assessment of user experiences. Specific monitoring probes are also deployed to accurately collect and analyze latency performance.


As per Fig. 1, the 5G-DT system relies on the 5G infrastructure available in the IT-Av site and it is composed by three main sub-systems: Smart Vehicle, Smart Camera and 5G Infrastructure. The experiment evaluation also includes processing interface connectivity, as will be later described in the following section

5G Driving Trainer architecture
Fig. 1 - 5G Driving Trainer architecture.

The 5G-DT relies on the 5GinFIRE infrastructure deployed in IT-Aveiro site, as communication support and virtualization infrastructure. Communications are twofold: in one way, the OBUs establish over IEEE802.11p/WAVE protocol to the available RSUs tunnelling notification messages from the VNF to the driver’s smartphone. In the other way, Smart Cameras send visual descriptors to the VNF 5G-DT instance for driving assessment. The data aggregation performed in the cloud, collects vehicle descriptors from all the involved SCs to create a “driver profile” and evaluate the counter-measures to unsafe driving styles. Once the profile is evaluated against common traffic rules and hazardous actions have been detected, custom notifications are immediately forwarded to the user.

Other driving suggestions have been considered, so that even experienced drivers can correct inefficient driving styles or reduce the environmental footprint. The ECB sources videos stream from the relative Full-HD camera and carries all the local
computations to extract vehicle appearance and behavioural features, e.g., speed, trajectory, driving anomalies, emergency braking. Such visual descriptors are then made available to the 5G IT-Av cloud services (through the relative VNF 5G-DT  instance), to create a driver profile by aggregating data from all involved SCs. The set of IPcam and ECB is available through the 5GinFIRE VPN access for experimenting processing load balancing and stream forwarding to the virtual function instance. The overall
network architecture is depicted in Fig. 2.

Fig. 2 - 5G-DT system architecture

The 5G-DT services are integrated with the 5GINFIRE network platform, using the OpenStack instances in the IT-Aveiro facility. To gather messages from the VNF instances, OBUs from IT-Av automotive testbed are used, including the IEEE 802.11p/WAVE and WIFI channels. The WAVE link is used for mobility communications from the “Smart Vehicle” to the infrastructure and WIFI is available for in-vehicle connectivity.

On the RSU side, pre-processed data and visual descriptors from SCs are forwarded to the VNF instance through in-house Ethernet links. Local IP addresses of IP cameras and ECBs are made available over the VPN during the experimentation and setup. Finally, the virtualized vehicle images are executed in the IT-Av Automotive cloud, using the OpenStack VIM platform.

The experimental deployment setup is depicted in Fig. 3.

Fig. 3 - IT Aveiro testbed facility and Smart Cameras locations (orange).


Contact Us

We're not around right now. But you can send us an email and we'll get back to you, asap.

Not readable? Change text. captcha txt

Start typing and press Enter to search

By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.