As a city or suburban driver, you know the frustration of standing in a long line of vehicles at a traffic light. You’re waiting at an intersection during rush hour. The red light seems to take forever to turn green until the first cars and trucks finally begin to move through the intersection. You follow at a slow pace, only to come to another standstill as the light turns red again. This cycle repeats itself over and over. You can’t help but wonder when you will ever get through. Finally, it’s your turn to move through the intersection! You make it a few hundred meters before the next light turns red. So yes, road rage is a very real thing, and the frustration seems to be growing every day.
Surely there must be a better way, right? Technical visionaries have been thinking about exactly that question for years and have come up with a solution: Intelligent Traffic Light Control Systems (ITLCs). ITLCs can monitor and analyze the traffic coming in and keep the timing of the traffic lights in sync so that things go as smoothly as possible. They communicate with other traffic lights along the route, preventing or reducing congestion.
There are a lot of big and small companies working on smart traffic light technology right now. In the Dutch city of Eindhoven, also called Europe’s center of innovation by Fortune.com, high-tech masterminds from around the world are flocking to meet current societal challenges with sophisticated new technology. It is here that ViNotion’s computer vision experts started working on real-time detection and tracking of traffic modalities when, in 2017, the first ITLCs hit the market. We’ve been talking to ViNotion’s CEO, Egbert, and sales manager, Azem, and their top-notch team of developers and engineers, about the challenges and technological advancements in artificially intelligent video interpretation, about smart sensors, and the future of intelligent traffic light control.
Traffic officials can’t do their jobs well at intersections because they don’t have a full picture of what’s going on
A group of Dutch traffic experts found it hard to solve traffic problems because traditional traffic light control systems only see heavy traffic and have trouble telling the difference between one or more bikes. Because pedestrians and cyclists weren’t factored in, the traffic authorities didn’t have a complete picture of the traffic situation. This made it challenging to effectively manage traffic at intersections. To find a solution, they enlisted the help of ViNotion’s knowledge and technology.
Ph.D. students from Eindhoven University of Technology (TU/e) joined ViNotion’s research team. These students have been looking into how to use Artificial Intelligence to automate the human visual system. Together, they made smart new traffic sensors with state-of-the-art video analytics software that is powered by AI. These sensors can work with inductive loop detectors on sidewalks and roads or even replace them. The system can easily notice both motorized vehicles and other road users, like cyclists and pedestrians, who are coming up behind it.
With these sensors in place, it should be possible to control traffic lights dynamically based on metadata output. So, ViNotion put up test models at busy intersections and on a bike path close to the Amsterdam Schiphol International Airport. The results were readily apparent. The results showed that the average amount of time cyclists had to wait was cut by 5 seconds, which meant that most cyclists didn’t have to stop. Since then, ViNotion has put its ViSense® AI video analytics technology and smart sensors to use in a number of places in Europe and the UK.
“Unlike other intelligent video surveillance systems that don’t catch cyclists and pedestrians or that rely on third-party devices like smartphones, our systems generate in-depth data on all types of traffic,” says Christopher, the Operations Manager at ViNotion. He continues, “ViSense® considers all traffic approaching an intersection and does not require personal tags or devices to track the individual traffic participants.” We measure velocity, travel direction, lane changes, and can even indicate abnormal behavior. If applicable, we add thresholds for warnings. The data we collect is sent right away to a traffic control application. In the Netherlands, the traffic control application is hosted in the cloud and the communication to this application is done via a broker called Urban Data Access Platform (UDAP). Through the UDAP, our versatile data output is exploited by the traffic light control application that sets off a traffic light schedule in the intelligent traffic light control system, ITLC for short, to move traffic across the intersection. Doing so controls actual traffic on the roads in real-time! “
Moreover, Azem adds to the conversation, “intelligent traffic light control systems have access to data and traffic plans from other traffic lights along the route.”
When an intersection creates a schedule and executes it, it can tell its downstream neighbors what traffic they should expect according to the plan. Smart traffic lights will work even better as more cars, trucks, and other vehicles, like emergency vehicles, use networked vehicle technology that lets them talk to each other and with infrastructure like traffic lights. If you imagine a future where all vehicles or all travelers are connected, then that could fundamentally change how we control traffic lights. Smart traffic lights wouldn’t have to wait for drivers to come within range of cameras. Instead, they could make decisions in advance based on information they get from drivers, truckers, and cyclists about their location, direction of travel, or even their entire planned route.
“We believe it will be a while before this becomes the norm,” Egbert replies when asked about it. Among other things, there are still many concerns about cybersecurity and data privacy in the realm of the connected cloud. ViNotion’s ViSense® system does not rely on connected third-party devices. With the convenient and, more importantly, flexible placement of our smart sensors along the road in the existing infrastructure, the behavior of each road user can already be monitored and analyzed in real time. By repeatedly placing a small array of intelligent sensors well in advance of an intersection, we can create sufficient time and distance for traffic participants to adjust their speed and lane position to form groups. This has proven particularly useful for prioritizing groups of road users, such as trucks or even bicyclists. “
Cars and trucks share roads with cyclists and pedestrians, sensed by intelligent traffic light control systems. Smart traffic lights recognize pedestrians at street corners and calculate how much time they will need to cross an intersection safely. ViNotion has a single sensor device that can read all road users coming up to an intersection. This device can automatically send GPS positions, speeds, and directions of travel to the traffic management application so that the Estimated Time of Arrival (ETA) can be calculated automatically.
” Let’s say you’re walking with a stroller or a walker. As soon as the system spots you, it knows how long it will take you to cross the street, explains Azem. That’s what the system tells the traffic light in advance. So, when you get close to the crossing, you get the green light right away. Plus, you can be sure you have enough time to cross. “
“Traffic lights have traditionally been geared toward motorized vehicles, but in our cities, we’ve moved toward more ways to get around, like walking, biking, public transportation, trains, and more,” he says. “ViSense® Video AI solutions for ITLC can detect, categorize, and analyze all road users to allow a calmer and safer traffic landscape while reducing pollutant emissions.” The main reasons for using ViSense® are to improve road safety and traffic flow, but the other benefits may make an even bigger difference. “
From induction loops to Artificial Intelligence.
With the number of cars on the road going up all the time, traffic jams have become a major problem. Commonly, we deploy traffic lights to control traffic flow at road intersections. Most traffic signal systems still use lead times and countdown timers to control traffic flow. Due to these fixed times, too often, traditional traffic control systems are just not able to handle unexpected traffic flows, resulting in traffic congestion. We need adaptive traffic light control systems that can watch traffic in real time and change the timing of the lights based on how busy the roads are. While induction loops appear to be a useful and widely used tool for intersection control, they are not without flaws. Performance-wise, induction loops only detect the presence of vehicles in their immediate vicinity. To find vehicles coming from more than one traffic lane, you need more than one induction loop. What’s more, maintaining inductive loops requires costly road repairs and disruptive lane closures. ViNotion has developed a remote sensing solution that does not require complex infrastructural changes or traffic-disturbing maintenance for deployment. ViNotion’s AI-powered sensors don’t just track where and how vehicles move; they also use advanced classification algorithms to tell the difference between different kinds of vehicles. Intelligent traffic light control systems can automatically give priority to emergency vehicles, public transit buses, and vulnerable road users by distinguishing between a truck and a public transit bus, or between cyclists and car traffic. As traffic and vehicles are evolving, our infrastructure must evolve too. With a better understanding of how vehicles, pedestrians, and bicycles move across crowded intersections, both in real-time and at different times of the day, road and traffic operators are now able to identify dangerous situations. It enables them to intervene to eliminate bottlenecks far beyond the capabilities of induction loops. Also, ViSense® gives space and infrastructure planners useful information that they can use to plan for the future of our intersections and roads.
Smart camera sensors and artificial intelligent analysis of content are paving the way for intelligent control of traffic lights in the future.
Initial advantages of the ViSense® system include comprehensive situational awareness of all traffic at an intersection and real-time connectivity. It allows us to know exactly how many vehicles, trucks, cyclists, pedestrians, or buses are passing through an intersection and how those road users are behaving. This enables our road and traffic authorities to optimize traffic flow in real time. When combined with smart sensors, ViNotion’s intelligent video analysis technology not only helps to cut down on travel and waiting times at traffic lights, but it also helps to improve safety. It also has a lot of historical information that can be used to plan road infrastructure and improve space and infrastructure to make our roads safer.
Intelligent traffic light control systems are gaining acceptance and use in the EU, but widespread implementation has not yet occurred. Expense may be a contributing factor. Depending on location and any necessary intersection improvements, ITLC system costs can vary significantly. The small ViSense® sensor units from ViNotion are easy to use for both new (ITLC) and existing TLCs.
Traffic experts think that driverless, self-driving cars will become more common in the next few decades. This is likely to make intelligent traffic control even more important, whether with traffic lights or other communication signals. We are moving toward greater connectivity in all areas, and that includes our traffic and transportation modes. If we are ever going to deploy autonomous vehicles on a larger scale, we will need to develop infrastructure that is able to deal with the governance of individual road users to safeguard the general interest and local policies.
Our roads are chronically jammed, and vehicles queue endlessly at intersections. At rush hour, long traffic jams are particularly bad.
So, the next time you get stuck in a red-light traffic jam—probably later tonight when you leave the office—turn up the volume on your podcast or guilty pleasure playlist and drift along, knowing that better days are on the horizon. ViNotion is one of the companies that is leading the way in using Artificial Intelligence to make traffic lights smarter. Machine-learning algorithms and smart sensors are well on their way to improving traffic flow, shortening wait times, and making intersections safer for everyone, including pedestrians and cyclists, in the near future.
Fun-Fact
Intelligent traffic lights can play an important role in the fight against climate change by combating time loss due to gridlocked traffic. A 2022 study by Juniper Research found that smart traffic management systems could reduce global emissions by 226 million tons by 2027.
