By Shane Nolan

AI has been the buzzword of the year, and many get caught up in the hype, dreaming of a future where artificial intelligence manages everything, including parking management. But let’s be realistic, expecting an autonomous system to manage parking and enforce compliance on its own is just not practical.

A fully automated model makes great marketing, but it’s crucial to remember that public acceptance is key in our industry. A staged deployment is necessary to ensure both operators and the public are comfortable with how new technology operates.

One of the biggest concerns we hear from parking directors is the risk of sending tickets to people who don’t deserve them. Can you imagine the backlash? Who in their right mind would stand up and say, “Sure, let’s just let the system handle all the tickets, no questions asked”? As industry providers, do we really want to push solutions that could put our clients on the unemployment line?

Case Study: The Evolution of License Plate Recognition

License plate recognition (LPR) technology has come a long way. We started with people manually capturing images of license plates using handheld devices. Those early systems really depended on human input to process and verify the data. Then, technology took a big step forward when optical character recognition (OCR) cameras were added to handheld devices and vehicles, allowing for automatic capture and recognition of license plates. This game-changer cut down manual data entry and made enforcement much more efficient.

These systems were improved to compare recognized license plates against pre-set business rules, like checking if a vehicle is parked legally or if it’s overstayed in a time-restricted zone. This integration of business logic made enforcement more automated and accurate.

Now, these systems can identify vehicle plates, compare them against the parking rules, and determine if the surrounding curb elements are a factor for each vehicle. The use case described here is when a camera detects a vehicle plate, confirms it has a valid payment or permit, and knows the vehicle is parked illegally in front of a fire hydrant. This scenario had to rely on experienced enforcement personnel to identify the conflicting information, or this infraction would have been missed as a paid plate. This level of logic also applies to enforcing disabled parking spaces on- and off-street.

The next big leap? We must find realistic ways to apply AI, not just for the sake of it, but also to assist human operators in filtering the volume of conditions and bringing exceptions to the front for further action. Imagine AI systems that don’t just recognize license plates but also recognize “vehicles as objects,” even with a number plate that is difficult to recognize due to obstructions. This exception can be presented to the operator for action and then continue to cross-reference multiple data sources to make smarter decisions.

AI’s Present Impact and Vision for the Future

We’re envisioning a future where these solutions not only recognize and compare data more accurately but also blend seamlessly with other urban infrastructure. But let’s be clear—the goal isn’t just advancing technology. It’s about improving the accuracy and efficiency of parking operations. We focus on supporting parking management and enforcement to create better urban experiences, not just winning the AI race.

Currently, the industry is using AI-driven License Plate Recognition (AI LPR) systems and Ticket-by-Mail solutions to automate enforcement with minimal human supervision and reduce the need for enforcement personnel to be exposed to potential safety risks. AI LPR cameras can improve the process of scanning vehicle license plates, detecting violations, and automating the process of sending tickets (Ticket-by-Mail) to registered vehicle owners without needing to place a paper ticket on the vehicle or potentially becoming engaged with the vehicle’s occupants.

Medford, Massachusetts, uses an award-winning parking enforcement system featuring AI-powered LPR and remote enforcement with Ticket-by-Mail to maintain an impressive 95% compliance rate. In Kitchener, Canada, AI LPR technology enforces parking rules during peak hours around schools when children arrive and leave. Authorities there have reported a more than 30% increase in parking tickets year over year since implementing the technology. These examples highlight a critical point: parking violations are far more prevalent than traditional methods can effectively monitor. AI has become a proven way to enforce these rules consistently, and the results speak for themselves.

And this is just one example of how AI can revolutionize parking enforcement and management, providing more efficient and consistent oversight. As technology advances, there is significant potential for further innovations that can enhance every aspect of parking management, including predictive analytics for optimizing parking resources.

The Framework of AI Progression in Parking: Assisted, Augmented, and Autonomous

Assisted AI supports human decision-making with enhanced insights and analytics. Augmented AI is more active and can suggest actions based on exceptions and data patterns. Autonomous AI operates independently, managing parking operations with minimal human intervention optimizing efficiency and resource use in real-time.

1. Assisted AI: a Tool to Support Parking Managers

In parking management, Assisted AI can revolutionize how we use data-driven insights to optimize operations with more informed decisions and better control over the process.

A practical example is AI-driven parking occupancy analytics, where AI analyzes data from cameras, sensors, and historical records to provide real-time insights into parking lot occupancy rates, trends, and usage patterns. Detailed reports and visual dashboards can help parking managers understand how their facilities are used, anticipate peak times, and adjust space allocation and staffing accordingly.

The key advantage of Assisted AI is that it empowers parking managers with enhanced visibility and predictive insights while leaving the final decisions in their hands. By identifying inefficiencies and suggesting areas for improvement, AI helps optimize operations without removing the human element.

2. Augmented AI: Enhanced Decision-Making with Actionable Insights

Unlike Assisted AI, which primarily supports human decision-making, Augmented AI takes a more proactive approach. It interprets complex data sets, identifies patterns, and proposes specific actions to optimize operations—with solutions your experts might not have immediately considered. Still, humans retain ultimate control over the final decisions.

Such tools can be particularly valuable in dynamic pricing strategies. For instance, an Augmented AI system can analyze real-time data on parking lot occupancy, local events, weather conditions, and historical trends to suggest optimal pricing models. If a major event is happening nearby, the AI might recommend increasing parking fees to capitalize on the higher demand while suggesting alternative parking allocations to manage the influx of vehicles more effectively. The AI doesn’t just present data; it actively recommends pricing and space management adjustments to revenue and improve customer satisfaction.

3. Autonomous AI: a New Generation of Tools for Minimal Human Input

AI-enhanced License Plate Recognition (LPR) cameras are a great example of semi-autonomous technology. These systems use advanced AI to scan and recognize vehicle license plates with impressive accuracy—up to 98% with today’s technology. When the confidence level of a read is high, it often eliminates the need for continual human oversight. This precision saves time for parking management staff, who no longer need to manually verify and record infractions.

However, it’s important to note that this is just the beginning. We can’t yet apply this as a blanket rule for all AI technology in parking management. While AI can partially handle the job, regular human checks are still necessary to ensure the logic remains sound.

Fully autonomous AI is meant to operate independently, processing data, making decisions, and acting in real-time. In current real-world applications, full autonomy often falls short. We’ve seen situations where human intervention was still needed to fix errors, like those caused by a simple software update. So, while the technology promises autonomy, the reality is that it often still requires human hands to step in when things go wrong.

Maybe full autonomy shouldn’t even be the ultimate goal. The goal should be to streamline operations, create efficiencies, improve safety, and meet customers’ needs today while anticipating what they might need tomorrow. As impressive as AI is, it’s here to support, not replace, the human element. While we continue to innovate and push the boundaries of what these systems can do, it’s important to stay grounded to provide reliable, efficient solutions that enhance the customer experience without being distracted by buzzwords.

Future Potential and Considerations for Implementation

AI-powered systems will become increasingly adept at managing parking resources without or with less human intervention. For example, AI could help identify areas to improve the parking operation or alert managers to potential problems, allowing them to be proactive versus always reactive.

In addition to operational efficiency, Autonomous AI holds significant promise for improving security, especially for enforcement officers. By integrating AI systems with local law enforcement databases, AI could identify keywords from incident reports—like gunshots or other disturbances—and alert parking enforcement officers to avoid certain areas.

While the legal aspects of inter-agency information sharing need careful consideration, limited access to real-time incident data could improve officer safety by helping them avoid dangerous situations. AI could also be integrated with 311 systems to alert officers to avoid areas where a public works response is required, further enhancing safety during enforcement.

Implementing any advanced parking management and enforcement system comes with legal and infrastructure challenges. For example, in Galveston, Texas, the law had to be changed to allow data sharing with the DMV to make Ticket-by-Mail remote enforcement possible. In Pittsburgh and Hallandale, local street codes were updated to legalize remote parking enforcement, as the old laws didn’t account for this new technology.

These cases show how laws must keep up with technological advancements. Infrastructure issues, like ensuring reliable data access across systems, are also important for making these AI systems work smoothly.

Public perception is another critical factor. While AI can enhance safety and efficiency, concerns about privacy, data security, and the potential for over-reliance on automated systems may exist. As we move toward a future where autonomous AI plays a more significant role in parking management, addressing these challenges will be crucial to fully benefit from this technology.

About the author: Shane Nolan, Director of Product Management, gtechna

Shane Nolan has a proven track record of success and is the Director of Product Management at gtechna. He is responsible for creating successful products that provide real business value to municipalities and their communities. With extensive expertise in parking management and technology solutions, he has led organizations to improve products for greater efficiency and usability, leading to successful outcomes.