Original article was published on Artificial Intelligence on Medium
Advanced traffic management: a powerful union of robotics, IoT and AI
As artificial intelligence (AI), the Internet of Things (IoT) and robotics develop, so do the ways they can help make city traffic less of a sweat.
Congestion, pollution, accidents, and imperfect infrastructure remain a problem for major cities. With the development of smaller urban areas, the same challenges become more widely spread.
Luckily, cities across the globe increasingly understand the need to invest in the latest technology to improve traffic conditions. We have seen dozens of daring initiatives from advanced development planning to dynamic management undertaken by local authorities in the past decade.
Projects that were launched as experiments as long ago as 2011 are now turning into best practices in advanced traffic management. Smart traffic lights, fully automated parking areas, and mobile device tracking have proven themselves highly efficient.
On the other hand, there are new challenges ahead, such as the growing use of autonomous vehicles, and these challenges need to be addressed now.
This article will guide you through the most advanced techniques in the aforementioned areas of improvement.
By making parking zones more efficient, municipalities can get the most out of the public spaсe, and edgy technology is gradually finding its way into this area.
If you as a driver know exactly where to go to safely drop your car off, you will easily avoid making circles around the block in search of an occasionally free slot. And when the network of devices connects meaningful data that’s then processed by AI, local authorities will be able to see patterns and use them in developing a transportation system.
Take automated valet parking. It was first introduced in 2018 by Stanley Robotics at the Lyon-Saint Exupéry Airport in France. The system creates an additional 50% of parking spaces and tangibly reduces CO2 emission levels.
If you travel by air from this airport, you can go to its website and reserve a parking spot. As you arrive, you leave your car in a box for the valet robot to pick it up and park it in a space-efficient manner. See the video below to understand how this works:
But robot-based automation can do more if you add in a system of ramps, shelves, slabs, and lifts. Connect this system to a computer, set AI to orchestrate it, and you will benefit from maximized space and happy car drivers who can book parking slots and get their cars on the fly by calling or using a mobile app.
Regardless of the robotization level of a parking area, there is one type of IoT device that alone can boost its efficiency: smart sensors. Such wireless sensors are typically mounted in the ground. They can monitor the situation at the parking lot in real-time, sending notifications to drivers when there are available slots.
There are two major alternatives to sensors: smart cameras and overhead radars/lidars. The former is cheaper but also less accurate, while the latter are much more accurate yet more expensive.
UK-based PayBySky offers a new approach to parking, making the process even more dynamic. They combine satellite vehicle positioning algorithms with a database of parking areas, including information like rules, time, and prices.
You install Skymeter, the device that makes your car visible to satellites and other connected devices around, into the OBD-II port. The system will pay for you once you’ve parked, and you are automatically charged on a quarterly basis.
And if you choose to park in a spot that’s illegal or obstructs traffic, the system will alert you.
Traffic light systems play a huge role in city infrastructure. Countries across the world have long worked on making them more reliable and manageable to reduce casualties and generally improve their citizens’ quality of life. This is where a combination of IoT and AI can make a big difference.
In 2018, a partnership of the city of Dallas and Ericsson resulted in the real-time adjustment of traffic signals at hundreds of intersections. They also used this system to prioritize buses on the road by targeted greenlight timing.
The authorities in Vienna announced the roll-out of a smart traffic light system in 2019. It was preceded by a testing period during which the engineers from the Institute of Computer Graphics and Vision at Graz University of Technology trained the AI.
The system comprises 200 lights equipped with small cameras that use an AI technique, computer vision, to recognize a pedestrian’s intention to cross the street. Once the intent has been identified, the light changes to green, giving priority to the pedestrian.
In addition, these lights have their own fault monitoring system. If anything goes wrong, it reports the situation immediately. This helps reduce and, in some situations, prevent those moments of street chaos that everyone hates.
Mobile device connectivity has been increasing rapidly in the past decade. By tracking devices across the city, municipalities can analyze traffic patterns to develop efficient public passenger flow and avoid bottlenecks that could otherwise lead to traffic jams and accidents.
Cities can locate connected mobile device users with network-based positioning algorithms that receive radio signals from base stations nearby. In 2019, EURASIP Journal on Wireless Communications and Networking published a study suggesting an interesting mobile device tracking algorithm: antenna, map and timing information–based tracking (AMT).
This approach is especially suitable for automated traffic monitoring as it does not require collaboration from mobile device users or modifications on the server. AMT tracks mobile devices using “enriched open map data, a mode of transportation estimator, and advanced route filtering on top of the mobile cellular topology and measurements.”
One of the world leaders in Smart City Initiatives, Denmark is already implementing a set of mobile tracking technologies. Copenhagen Connecting exemplifies how this can help optimize traffic and reduce congestion.
In 2018, the City Council of Portland, Oregon, unanimously voted for an agreement between the Portland Bureau of Transportation, TriMet and the regional government Metro to fund testing of Replica, a program developed by Alphabet’s subsidiary Sidewalk Labs.
This tool uses precise location data from smartphones and data from the U.S. Census Bureau to create a digital twin of the Portland metro area. According to Portland’s transportation planners, this is a great way to “dramatically improve their ability to understand and monitor how people use streets, sidewalks and transit.”
Autonomous vehicle systems
Driverless cars are gradually gaining a foothold in our cities. They bring multiple benefits, such as reduction of accidents and lower congestion, but they also come with new challenges.
Here is one scenario: according to a report published in Transport Policy in 2019, driverless cars might even increase congestion if they keep circling around to avoid paying parking fees, which are rather high in major cities.
Self-driving vehicles need to be managed. As Siemens Mobility CEO Marcus Welz put it in an interview at CES 2020 in Las Vegas: “Frankly, thousands of self-driving vehicles are still thousands of vehicles in a traffic jam whether there’s a human driver or not.”
What happens if people without licenses, who have never owned a car and did not even plan to, increasingly buy autonomous cars due to their growing availability? The load on the roads will reach a historical maximum and our cities will freeze in traffic jams.
Autonomous vehicles are IoT devices, and they integrate into the bigger traffic management system. This provides an opportunity to create a smart management system for such vehicles. Municipalities should think about it now before driverless cars flood the streets.
There are two directions that public and private sectors should work together in:
- The intelligent use of vehicle-to-infrastructure (V2I) systems comprising road sensors and connected road signs that speak with autonomous vehicles. Local and national authorities should decide what approach works better in each case: dedicated short-range communications (DSRC) or cellular vehicle-to-everything (C-V2X) systems.
- Encouraging shared autonomous mobility (SAM) by creating an interconnected network of static and mobile sensors and mobile apps.
It is interesting that scientists have been developing carpooling with autonomous vehicles in mind for a few years now.
In 2016, MIT published a study led by Professor Daniela Rus of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL). The researchers developed an algorithm that reroutes cars in response to incoming requests and sends idle cars to the areas where the demand is high at the moment, all in real-time.
Professor Rus specified that “the system is particularly suited to autonomous cars, since it can continuously reroute vehicles based on real-time requests.”
As Smart City techniques mature, they shape the expectations of citizens. The UK government, for example, is already working on smart parking standards, and compliance with some Smart City standards, such as ISO 37122, is only achievable by municipalities that have introduced smart parking.
The technologies behind the latest advancements in traffic management are complex and require expertise to not only introduce them but also to even make sense of how they work. Our team at Intetics Inc. has this expertise and will be happy to leverage it to your benefit. Let’s see how we can help.
Featured images by freepik.com