Data fusion for smarter, safer intersections
This article was originally published in Traffic Technology International magazine (June 2024). You can preview the original version here (click).
The vision of the US National Road Safety Strategy (NRSS) is simple: ‘that every trip of every kind on our nation’s streets, roads, and highways should end with a safe arrival.’(1) A sentiment not restricted to the US alone finds itself at the heart of international government politics. According to the US Department of Transportation, about half of all traffic injuries in the USA occur at intersections. The safety of Vulnerable Road Users (VRUs), like pedestrians and cyclists, is a high priority for the safety strategy (2).
Reducing crash-related fatal incidents is also on the agenda of city planners(3). While delivering truly autonomous driving and smart city infrastructure is the ultimate goal, there is a quiet revolution going on in our towns and cities to replace traditional detection technologies with more advanced radar and camera sensor solutions.
smartmicro intends to make roads safer for all road users. But how are traffic fatalities and serious injuries prevented through intersection management?
Today, four detectors independently manage the directions of a well-equipped signalized intersection, often resulting in blind spots at the inner part of the intersection and leaving movements across the intersection difficult to monitor. If VRUs remain unseen, motorized traffic cannot be controlled in real-time, allowing safe passage.
A smart intersection, including multiple sensors and an edge computer running sensor fusion software, provides a full overview, covering all approaches and the inner, previously less monitored intersection zone, tracking the position, velocity, and class of all objects in real-time. The communication between the sensors enables seamless tracking of all traffic participants - from a small child to a large truck - throughout the entire intersection. Utilizing data fusion between the integrated sensors, the perception of the environment is improved immensely. A rich set of collated data helps to identify potential safety hazards, triggering appropriate safety measures in real-time, such as flashing warning signals, traffic light adjustments, or messages to the drivers of connected vehicles via V2X communication. This means that vehicles can literally “see around the corner” in all directions of the intersection, knowing the traffic situation even before they arrive. This gives smart sensors at intersections a high value, especially during the time when vehicles connected to each other are still rare.
Additionally, a combination of radar sensors with cameras, along with data fusion – a hybrid sensor system – appears most promising, delivering reliable data under all weather and lighting conditions. Radar sensors uniquely provide true speed (Doppler) measurement, can deliver 4D data, have the longest range, and are maintenance-free.
The benefits of cameras lie in the visualization of a traffic scenario and the advanced classification of traffic objects. In the automotive sensor world, combining radar sensors and cameras has been the best practice choice for many years. The full system offers fail-safe features, such as blind and rain detection, misalignment reporting of the sensors, and interference suppression, which ensures dependable operation.
The main advantages of intersection management solutions are, firstly, the enhanced safety for vulnerable road users. Especially pedestrians and cyclists will be better protected, as they are reported to connected vehicles, thus reducing accidents. Secondly, broadcasted data can be fused with the vehicle’s onboard sensors (collective perception).
The signalized intersection of the future will be connected and smart, smarter than today, no question. The aim is to make roads safer – to fulfill vision zero – and to optimize traffic flow for not only but also reduced travel times, contributing to greener cities.