It does not matter if you are in the driver’s or passenger’s seat. We’ve all been through the same situation. The light that never opens. The endless queues of cars that stretch for miles. The deafening horns.
Traffic jams are a plague in our lives. But with so much technology at our disposal, why do we continue to deal with them so obsolete?
Although our means of transportation have evolved a lot in recent years, our traffic management systems have been having difficulty tracking the growth in the number of vehicles.
Flood containment measures often fail to respond to sudden changes in road or weather conditions. Not to mention the many traffic lights that still work with out-of-sync timers, preventing traffic from flowing normally.
By 2015, there were about 1.3 billion motor vehicles worldwide and with the rapid growth of emerging economies, that number is expected to rise to more than 2 billion by 2040. Even with new avenues, this increasing volume of traffic could quickly exceed the capacity of our road networks, especially in cities.
But could the combination of new communication technologies with the power of artificial intelligence (AI), which allow for large amounts of data to be processed in real time, be the solution to this problem?
Average speed too low
While many see autonomous vehicles as the outlet for traffic jams – since robots can be taught not only to drive less imprecisely, but also to react faster than human drivers – it will still take at least two decades before they begin to have a significant impact.
Until then, road agencies and urban planners will have to deal with an increasingly complicated mix of human, semi-autonomous, and autonomous drivers. Keeping them all moving will require that traffic management systems respond and adapt instantly.
In Bengaluru (Bangalore’s new official name), in India, a city that frequently records long traffic jams and where the average speed on some rush hour roads is only 4 km / h, the technology giant Siemens Mobility has created a prototype of a monitoring system that uses AI through security cameras scattered along the tracks.
Cameras identify the number of vehicles in real time and transmit the information to a control center, where algorithms calculate traffic density. From this data, the system changes the cadence of traffic lights.
But this requires data. Lots of data. Fortunately, they are plentiful. There is a wealth of information from traffic monitoring systems, road infrastructure, cars and drivers through cell phones.
Millions of cameras are scattered along our roads as moving vehicles induce small electrical currents into metal devices hidden under the asphalt, providing more information on traffic conditions. Drivers can send instant updates on delays thanks to the navigation software they use on their smartphones and their cars.
High data volume
Some of these monitoring technologies – such as induction loop – exist since the 1960s, while others, such as cameras capable of tracking traffic and reading boards, are newer. The challenge is to optimize all this information and make it useful.
“Since Isaac Newton, we have tried to influence the world through the construction of mathematical models,” says Gabor Orosz, an associate professor of engineering at the University of Michigan in the United States. “If we have the data, we can come up with solutions. The same applies to traffic.”
Today, there are ongoing initiatives to leverage AI’s ability to understand large amounts of information and change the way we move around our cities.
Recently, researchers at the Alan Turing Institute in London and the Toyota Mobility Foundation in Japan have launched a joint project that aims to improve traffic management systems through the use of artificial intelligence.
Scientists simulate scenarios that become increasingly complex, helping algorithms learn how to predict changes in traffic. Although the system is still in the testing phase, the expectation is that it can be used soon in the real world.
“With deep machine learning, we can improve predictability,” says William Chernicoff, head of research and innovation at the Toyota Mobility Foundation. “Those responsible for urban mobility can then make faster and more efficient decisions about traffic light time, suggested routes for system users and capacity allocation.”
In Pittsburgh, in the United States, researchers are already working with city managers on a similar initiative that has run the city since 2012. A traffic control system developed by researchers at the Institute of Robotics at Carnegie Mellon University was deployed throughout the city for a company called Rapid Flow Tech.
Its technology, Surtrac, is being used at 50 intersections in Pittsburgh and since its launch has reduced standby time by as much as 40 percent, according to the company. The company also claims that travel times in the city have dropped by 25%. The emissions of pollutant gases also registered a fall, of up to 20%.
The system uses video cameras to automatically detect the number of road users, including pedestrians and types of vehicles that are at a crossroads. The software, powered by AI, processes this information second by second to find the best way to ensure traffic flow by resynchronizing traffic lights, depending on what is ideal for keeping traffic moving. Decisions can be made autonomously or shared with other crosses to help them understand what is happening.
As vehicles become more connected with the help of cell phones and other wireless technologies, they also help power such systems with more information. In the future, according to Griffin Schultz, CEO of Rapid Flow, connected vehicles will be able to share information about their speed, driver behavior and even possible flaws in the surrounding infrastructure.
“At the moment, we are only learning, but this will be much more common in the future,” he predicts. “It’s not just about cars, but this technology will help all types of road users in a multimodal transport society.”
Elsewhere in the world, intelligent infrastructure has been helping transport networks become more connected. Siemens Mobility is operating in cities and counties around the world to identify patterns of movement in an attempt to enhance everyone’s experience on the street.
“There are real projects around the world and their applications are constantly expanding,” says Markus Schlitt, director of intelligent traffic systems for the company.
“In the cities of the future, traffic will be so complex that without artificial intelligence we would be stuck in traffic,” Schlitt says. “By using the data, we can identify patterns that would not be seen without AI.” Through this continuous learning, we can constantly update traffic patterns and thus vehicle flow, resulting in less waiting time and fewer emissions. ”
Bikes also in focus
In Hagen, Germany, artificial intelligence is being used to optimize traffic light control and reduce waiting time at a junction. The simulations indicate that the system can reduce waiting times at traffic lights by up to 47% compared to pre-programmed ones.
But it’s not just the drivers who are benefiting from the use of AI. Siemens Mobility is operating a fleet of 1,400 electric bicycles in Lisbon, using machine learning to analyze various data sources, such as weather, to forecast future demand at each of the 140 rental stations.
In this way, the company can guarantee not only the availability of bicycles, but also free spaces for when they are returned. Such data is used in conjunction with the latest traffic information to assist in replacing the bibi jacks at rental stations and to set the ideal caster for maintenance technicians.
“This not only reduces operating costs but also improves the end-user experience,” says Schlitt. “So if you need to move in Lisbon, you can be sure there will always be an electric bicycle available at the stations.”
However brilliant the technology may be, we can not rely on it alone. Mischa Dohler, from the IT department at King’s College London and co-founder of traffic monitoring technology company Worldsensing, has been testing artificial intelligence and machine learning in Bogotá, Colombia.
He says the technology has already produced great results, such as using signs and traffic signs to redirect traffic when there is an accident, reduce traffic jams and decrease the time that drivers spend looking for parking spaces.
But he says that while AI is helping to make this kind of adaptive transport network possible, the human element remains important. He calls this “explainable IA planning”. Basically, it allows humans to make decisions together with AI or adapt if something goes wrong. Although they are intellectually and technically capable, drivers themselves must be open to the idea that their traffic systems are controlled by computers.
“When cities rely on algorithms to implement policies, that policy is overshadowed by computing,” says Jed Carter, editor of online magazine Moving World. “It becomes even more difficult for citizens to understand why they were redirected, photographed or retained when the reasons for such actions are related to computer code.”
But the implementation of intelligent technologies on the roads is not only to avoid traffic jams. Mark Nicholson of Vivacity Labs, who coordinated a UK government-backed project deploying smart traffic lights in Milton Keynes, England, says the latest technologies have many other benefits.
The cost is one of them – as technology takes on more prominence in traffic management, the less need for human intervention in basic tasks such as tracking traffic cameras.
Automated systems can also increasingly differentiate a large number of road users. That way, depending on the circumstances, cyclists, buses or emergency vehicles can be prioritized.
By keeping traffic flowing, they also reduce the power consumption of slow or stationary vehicles, improving air quality and benefiting the environment. Lastly, they help drivers find faster parking spaces, favoring their productivity.
“With automation, we can focus on what’s most important,” says Nicholson. He exemplifies: “Things like improving air quality near a school, avoiding the passage of trucks or other heavy vehicles, planning where we are going to build a new diversion or practical issues, on how we will redirect traffic after an accident.”
For Nicholson, the real benefit of technology is to enable humans to maximize their potential. But how? Preventing us from wasting time in daily, tedious traffic control activities, he says. So, with the help of AI, we can focus on what we are best at, as in situations that require adaptive thinking and creative solutions.
The results of the Milton Keynes project are promising. City-wide smart cameras capable of identifying and classifying users and vehicles have enabled extremely accurate data, providing urban planners and authorities with information on peak traffic times, the most popular routes, and the availability of parking spaces.
Vivacity has installed 411 of its smart traffic cameras at the main junctions in Milton Keynes. In addition to counting and sorting users, sensors can measure the time it takes for vehicles to move between intersections and provide real-time photos to help with future planning.
The company sends the data to a machine learning model that learns typical daily patterns and combines this with the way traffic responds to transient changes in the road network. The system evolves and adapts over time, improving its predictive power and minimizing the level of human intervention required. It also provides historical and real-time data, as well as predicting daytime traffic.
As a result, it was able to predict traffic conditions 15 minutes in advance with 89%.
“The system is not only helping citizens check the availability of parking spaces in real time, but it also lays the groundwork for future connected and autonomous transport technologies in Milton Keynes,” says Nicholson.
What seems to be clear is that giving AI the green light will allow us to keep moving forward.
“This is just the beginning – we do not fully capitalize on the capabilities and benefits of AI,” adds Markus Schlitt of Seimens Mobility.
Translated by: Google Translate
Original content available here.