Get My Parking: Breaking Through Silos: Privacy and Data Concerns in Parking AI

We sat down with Aman Singh, our VP of Tech and resident parking AI expert as a part of the series of indParking Industry Yearbook about the true potential of intelligent parking systems, the un-siloing and standardizing of parking data, and the limitless possibilities that come with artificial intelligence.

In an era where we’re finding an infinite number of ways to involve artificial intelligence in our daily lives, even the simple process of parking is undergoing a transformative evolution. From smart sensors optimizing urban spaces to automated valet services, AI is reshaping the way we navigate and utilize parking facilities.

The application of AI in parking has been limited probably because of the nature of the industry, which in no way indicates the limits of what it can do for parking. So far, we’ve been using AI mostly for access control, to identify patterns in plate misreads, and to increase LPR accuracy.

AI can do way more than just LPR for the parking industry, but there are a few hurdles we need to overcome before we can use it to its full potential.

AI and Parking Data

Above all, AI needs data, to identify patterns and behaviors to make the right suggestions. As with most issues that concern data, there are a few questions that industry experts are still grappling with - how much data do you collect and how? Where do we store this data? What is the granularity and variety of data, and how much do we discard? How much do we share?

These questions aren’t easy to answer, so let’s look at an example.

Let’s take Manhattan, New York. Parking can get quite nightmarish in this area, especially during rush hour. If we had all the data from every parking location in Manhattan, we could do so much more in terms of traffic management by redirecting people to the right parking locations and taking cars off the road.

Today, parkers get occupancy data through operator apps, which only report on the locations owned by individual operators. AI can be used to identify empty spaces and inform the user whether there’s a chance to get a free spot.

Of course, for this information to be truly useful and accurate, the parker needs to know the total number of all the parking spots available in a certain area, regardless of who the operator is. The full picture can only be seen if data from all locations is available to the public and to all parking platforms. However multiple operators run the parking lots in the city, so the data isn’t shareable or public.

This is a limitation of any AI parking system in terms of what it can know and recommend. Once the data has been made available, AI programs can recommend the best and most affordable spots to drivers.

While AI in parking is a hot new idea with some great results, we’re only still scratching the surface because the data is siloed. Operators do not or cannot share their parking data because of limitations of their data infrastructure or because of their privacy rules.

There would be many benefits to having a common parking data ecosystem. With the right information, AI could elevate dynamic pricing help operators compete better, and streamline business processes and profits. Unfortunately, this is something that the industry has to come together to figure out.

One of the biggest challenges to unsiloing data is figuring out the integration and interoperability of multiple tech stacks. We have hundreds of companies in parking who are all working on solving the same problems. They all have their ways of working with data - how will this data flow into each other’s buckets? How do we make integrations seamless?

A step in the right direction is the Alliance of Parking Data Standards (APDS), which is a non-profit organization that helps develop, promote, manage, and maintain a uniform global standard to share parking data between platforms. It is a consensus-built international standard, establishing a common language for data elements and definitions in parking and mobility. The APDS is a great way to facilitate seamless integration, compatibility, and communication between parking entities, the automotive industry, IT developers, map/app providers etc.

AI and Parking Security

Taking parking data out of its silos and publicizing comes with its complications. Mishandling license plate data can be a threat to individual privacy. While data is secure in an encrypted cloud storage structure, it may not be safe in physical databases and devices at the location.

Apps that track license plate information can be misused to track someone’s location if the right rules and regulations aren’t in place. There was an incident where a woman’s ex-partner kept tabs on her whereabouts using a public website that allowed him to track every instance where her plate number had been read by an LPR camera.

While we should push for data to be publicized, there should be reinforcements in place that prevent the public from misusing sensitive information. The data should also be protected from possible leaks, since license plates, credit card information, and personal details like addresses are all collected and linked in a parking ecosystem.

However, AI can also help operators amp up their security measures. There are smart cameras available that offer anomaly detection, creating alerts on human movement, fire and noise. These can alert support staff to get rid of squatters or other disturbances in closed parking spaces like garages. This information can be relayed to customers through their apps, along with AI suggestions for alternate locations to park.

These cameras also help us track who gets in through pedestrian doors. It adds a layer of protection to make sure only authorized people have access to parking locations. With the help of AI, centralized management can be streamlined so that one control center can monitor more than 40 locations, offering a layered approach to how data is processed and alerts are sent out.

AI can Improve Parking Operations
If used right, AI can help parking businesses sell, manage and operate better.

Close tickets faster

AI can streamline customer support - we’ve seen great results with it at GMP. Our AI helps support teams fix and close sessions faster. The process involves pattern matching, string matching, and building heuristics based on parking behavior and hardware behavior. This helps us bring in large amounts of data which we convert to intelligence, which we use in real time to process and close sessions. For example, if a parker is stuck at an exit lane, the system can recommend five possible methods of exit to the support executive based on the traffic and pedestrian behavior.

Setting up AI chatbots on parking websites and apps with rule-based solution sets and operational procedures can reduce the time taken to receive and close support tickets.

No more misreads

The AI also helps us close sessions where the camera misreads the plate and the gate doesn’t open. Based on LP images from the user’s history, AI corrects the misread, verifies the number plate and opens the gate. This process is invisible to the user and happens superfast - the system repairs the misread using AI-assist and the gates open for the parker.

Push the right product

We’re also trying to figure out how much information can be derived from a parker’s actions. Deductions based on these logics help us improve space designation and solve customer complaints faster.

This also helps us market the right parking product to the right user. For example, the AI can deduce from a parker’s history or location whether they’re likely to buy a monthly permit or a temporary reservation. AI can use the parker’s location information to push better parking offers or validations created with retail partners, which can help operators earn more.

A system that adapts to earn more

Unlike e-commerce or other industries, the problems in the mobility and parking industry are very physical. There’s likely a car with a strange license plate that your camera can’t read.

Or there could be a thunderstorm that can change your whole day - the plates become harder to read, more cars come in to get off the road and into shelter. Your parking system’s ability to identify and adapt to these physical factors determines how much revenue you make.

A clever AI system can recognize physical data points like the weather or increased traffic and adjust the prices accordingly. AI-assisted LPR can make sure all the number plates are read based on their history, even if the cameras are fogged up. If the system has access to public data from traffic cameras and road sensors, it can make even better administrative decisions. With the right AI in place, a thunderstorm could earn you double the revenue.

AI and Self-driving Cars

The world has been talking about self-driving cars for a while now. The questions for the parking industry are: what would a self-driving car do when it’s not on the roads? How would it know how and where to park? How much information is available to the car to know what the best lot is nearby with the right pricing and free space?

There’s a lot of discussion about the car-as-a-wallet concept, where the payment method is linked to the car and its license plate instead of the user. The car is detected with LPR/Fastag/Bluetooth and the payment is instantaneous and automatic. But for the car to make the right decisions and end up in the right spot, the ecosystem needs to integrate several stakeholders and their databases. We have a long way to go with this concept, but it’s an interesting space to watch out for.

AI is here to stay

Looking ahead, the future of AI in parking holds exciting possibilities. Continued advancements in machine learning, sensor technologies, and connectivity are likely to fuel further innovation.

We can anticipate more sophisticated AI applications, increased automation, and enhanced user experiences as the parking industry continues to evolve in tandem with the digital age. The mark that AI is making on parking is not just a technological advancement; it's a transformative journey towards a more intelligent, efficient, and user-friendly parking ecosystem.

Read more exclusive interviews and articles written by industry experts in the Parking Industry Yearbook 2024 - grab your free copy now!

About Get My Parking

Founded in 2015, Get My Parking is an award-winning startup that has grown to a team of 150+ members across five continents. It provides essential technology to parking operators – white-label parking apps, an IoT gate kit to retrofit old parking gates for modern capabilities, and an interoperable cloud platform that enables centralized and digital operations across a network of parking lots. The startup is spearheading new trends like EV charging, connected cars, and shared mobility hubs on the parking real estate with plug-n-play API integrations. The GMP platform has been deployed across 3000 parking lots across the world and has processed more than a 100 million transactions till date.