Jan 31, 2021Urban transport systems are fraught with issues, but in many large, urban areas, none prove to be more problematic than traffic congestion. But traffic congestion isn't just a problem on its own—it can lead to larger, more troubling issues that can plague entire cities. Moreover, congestion often starts with, and is often caused by, poor parking.
A study conducted by Ali et al. suggests that, in the center of large cities, about 10 percent of traffic circulation can be traced back to parking issues—drivers cruising around on the hunt for a convenient spot, or for any spot at all. The study suggests that, on average, drivers can often spend up to 20 minutes searching for an available parking space. This can lead to traffic jams, higher energy consumption, increased air pollution, and more delays and accidents.
This is a big issue not only for those who drive, but also for business owners in large, urban centers. Approximately 40 percent of drivers surveyed in a study said they choose not to visit brick-and-mortar stores due to the difficulty of finding a parking space. With research predicting that around 68 percent of the world's population will live in urban areas by 2050, this problem needs to be solved sooner rather than later.
So what's the solution? According to Ali, it's necessary to develop a parking space availability prediction system that can inform drivers in advance about the availability and occupancy rates of parking spots based on day, location and hour. In short, by implementing the Internet of Things, alongside sensor networks and cloud technology, Ali et al. argue that a specific framework of their design, which is based on a deep long short-term memory network to predict the availability of parking spaces, could help to eliminate the issue of traffic congestion rooted in parking availability problems.
Can the IoT Solve the Parking Problem?
According to the study, the Internet of Things and deep learning can be used in the planning of smart cities, which could gradually tackle urban mobility problems, as well as help provide a sustainable infrastructure economically, ecologically and socially. On average, drivers spend about 3.5 to 14 minutes searching for an available parking spot, wasting their time and causing significant traffic congestion.
Ali et al. proposed a framework based on a deep long short-term memory network (an artificial recurrent neural network architecture employed in the deep learning field) to predict available parking spots using the Internet of Things—in practice, this involved the use of Birmingham parking sensor datasets to evaluate the performance of deep long short-term memory networks. Then, three types of experiments were performed in order to predict the availability of free parking spaces.
To perform this study, researchers adopted the architecture of smart car parking systems to gather car parking data through a variety of parking sensor networks deployed at various parking locations. Aggregated sensors collected this from different locations in the cloud, and then deep learning techniques were used to analyze this data to locate available free parking spaces through the sensors' network.
The idea behind this approach was to use sensor data to better understand available parking and then predict availability based on three experimental factors: the time of day, the day of the week and the location. With the help of the proposed system, which employs a deep LSTM network to predict the availability of car parking locations on a specific day of the week within a given time slot, drivers would be able to locate a parking space from any location at any time, thereby avoiding the endless idling, excess air pollution, traffic congestion and frustration.
How the IoT Can Improve Parking Conditions
Ultimately, according to the study, the experimental results showed that the proposed model outperformed other state-of-the-art prediction models. In the parking location experiment, the system predicted the availability of free parking spots in a given time at more than 90 percent accuracy; in the day experiments, the system reliably predicted parking lot occupancy on seven days from Friday to Thursday; and finally, it also reliably predicted the hourly parking lot occupancy from 8 AM to 5 PM. Ultimately, there is good reason to believe that it can help drivers find free car parking space near their destination—it can save time and energy consumption by efficiently and accurately predicting the available car parking space.
In the future, this type of predictive technology could extend into the framework of planning and building smart cities—in other words, smart parking based on the Internet of Things, cloud technology and predictive software might be a fundamental part of city planning. It might even involve more identifying factors like parking supervision, vehicle registration, vehicle tracking, identification and more. Ideally, smart parking will also evolve into real-time availability tracking rather than just predictive technology. Ultimately, all of these factors are thought to lead toward more efficient, more effective parking.
In a real-world setting, parking conditions can be vastly improved by the IoT. With predictive technology, and eventually with real-time notifications for available parking, it is possible that free, available parking will cease to be an issue in metro areas. With tools such as these, air pollution could be decreased, fewer traffic jams might result from parking problems, and congestion in cities could be reduced dramatically.
Hanna Marcus is a content writer for Do Supply Inc.. After earning a degree in journalism from the University of Florida and spending a few years in newsrooms across the state of Florida, she found her niche in content writing. Currently, Hanna specializes in several niches but enjoys writing about technology, artificial intelligence, the future of the IoT, and more.
