Adaptive traffic control with smart cameras in Varazdin
Improving traffic flow is a very important topic for cities as the number of road users keeps growing. To be able to get more out of the current road capacity, municipalities are turning to smart technologies like traffic detection and traffic control to improve the situation.
However, making significant improvements to traffic flow has not been possible until recently. For a long time, inductive loop technology was the only way to manage traffic adaptively but it could only do so much. Due to the high costs of inductive loops, only one was usually installed per lane meaning the traffic controller would only know that a car has recently passed in a specific point, with the use of current zone presence being limited to a very small area. On the other hand, with FLOW devices, you can utilize UDP sinks, namely the object sink which can tell the traffic controller the exact number of cars and their class that are currently present in each traffic lane giving the much-needed detail for much more efficient adaptive traffic control.
In addition to communicating the insights for real-time traffic controller decisions on the edge, the same camera can act as a multi-purpose sensor whose edge-extracted data can be pushed to or requested by 3rd party applications. For example, it can provide information to traffic management or decision support systems.
In Varaždin the municipality wanted to improve traffic on a busy 4-arm intersection next to a shopping mall and a factory that also connects to the highway. For this purpose, they have asked our partner TRAF SIGNAL to implement adaptive traffic control.
Staff training in Maribor traffic monitoring center
The municipality wanted to implement adaptive traffic control in general but did not have a specific requirement for how it should be done. On top of being able to manage the traffic adaptively to improve the traffic flow, the municipality was offered that the solution could gather traffic statistics such as vehicle counts for each direction and distribution of the traffic in terms of the vehicle types to better understand the traffic composition at this intersection. This was a free added value for them and therefore it was also implemented. The solution also had to be GDPR compliant as this was a deployment in an EU member country.
The last challenge was that this junction was further from the city center so there was no internet cabling at the location, therefore, LTE had to be supported in order to be able to send the real-time data to TRAF SIGNAL traffic management platform via a secured VPN.
To fulfill all of these quite demanding requirements in one solution, TRAF SIGNAL has decided to use TrafficCameras by DataFromSky with built-in FLOW video analytics specialized for traffic data gathering.
Traffic junction in Varaždin with adaptive traffic control using TrafficCameras
DataFromSky solution - TrafficCameras
The solution was to install smart TrafficCameras with built-in video analytics on the traffic light poles, one for each of the intersection arms. The cameras were able to detect incoming vehicles from a 70-meter distance allowing for responsive traffic signal switching without vehicles needing to stop when possible. The cameras were directly wired to the traffic controller using Ethernet Switch for local communication of the real-time data using UDP sinks. The system also had smart switch to Wi-Fi installed in case the specially secured cabinet door was open.
The cameras were installed on light poles by TRAF SIGNAL which was made easier thanks to detachable junction boxes. The evaluation of vehicle trajectories extracted from the video streams was done by the FLOW framework running directly inside the smart TrafficCameras. The video analytics were also remotely set up by them including the zones and sinks for adaptive traffic control as well as the setup for gathering various traffic statistics.
Example of a video analytics set up with visualized pedestrian trajectories
Our partner TRAF SIGNAL used its extensive expertise to set up and optimize the adaptive traffic control based on the real-time FLOW traffic data in two approaches depending on the traffic currently present at the junction. In the first one, the traffic controller triggered a green light for the given direction when just a single car was present at the intersection. In the second one, when the traffic levels increased the busier traffic lanes were given longer green light periods adaptively. After the full installation of the system, the traffic flow improved by 30% compared to non-adaptive traffic control.
FLOW dashboard with real-time traffic statistics – all aggregation is done by TrafficCamera
The TrafficCameras also gathered traffic statistics such as the category distribution of traffic and the number of vehicles from each direction. If the council wanted it would have also been possible to measure the approximate speed of vehicles approaching the intersection as well for all of the 4x directions.
The operating expenses of this system were significantly reduced thanks to the solution having only one LTE modem and communicating this data at longer intervals instead of streaming it continuously. One SIM card was used for the whole intersection and if required by the municipality, it would be even possible to push a whole RTSP stream with this setup. The edge processing of the data also ensured GDPR compliance. Moreover, only aggregated data was sent to the TRAF SIGNAL database with an interface that the city can access which further reduced data communication needs. Lastly, the TRAF SIGNAL system also generated low-bandwidth JPEG images for real-time previews of the cameras on their platform.
The SmartCameras provided much better long-term value, more functionality, and easier installation compared to induction loops which were considered the closest alternative solution. Continuous 24/7 monitoring and data gathering of detailed traffic statistics was achieved. A cost-effective video-analytic traffic monitoring setup was deployed thanks to the solution being camera-based, thanks to smart data usage, and most importantly thanks to the multi-purpose nature of the FLOW framework. A fully automated traffic control system was implemented taking into account traffic from each direction as well as the duration of the presence of vehicles in each of the waiting lanes and even the number of cars passing through the intersection for each green signal. The TrafficCameras are now gathering both real-time and long-term statistics at the same time about various elements of the traffic flow. The whole system has been running since October 2021 with only minor adjustments.
Want to read more? Go back to the BLOG.
- Real-time monitoring of 4x junction arms
- Collection of vehicle presence in zones for adaptive traffic control
- Direct communication with the traffic controller
- Gathering of traffic statistics and sending them to the data aggregation platform
- GDPR compliance
- Data connectivity
- 4x TrafficCameras installed on traffic light poles covering all 4x arms of the intersection
- Remote access via LTE, low FLOW data consumption setting
- Multipurpose FLOW framework for real-time adaptive traffic data as well as long-term traffic statistics gathering
- FLOW UDP protocol for communication with traffic controller
- 24/7 real-time adaptive traffic control and traffic monitoring
- Non-intrusive nature and fast deployment of the solution
- Remote and on-demand addition of new analytics features / traffic metrics thanks to FLOW framework
- Smooth integration with the traffic controller (direct data communication)