If there is one thing that drivers of the 1.5 billion vehicles on the world's roads have in common, it is gasoline. Driven by gas-powered vehicles' persistent need to refuel, fuel retailing is an $860 billion (USD) industry globally. In the United States, approximately 80% of its 145,000 gas stations are found in convenience store forecourts, according to the American Petroleum Institute (API) and the National Association of Convenience Stores (NACS). For consumers, filling the tank at a forecourt gas pump may be an experience they will repeat more than 10,000 times in a lifetime.
Akhileshwar Padala, who heads engineering at Allied Electronics, spends a lot of time thinking about how to improve the forecourt experience. As a well-respected industry expert in the highly specialized field of fuel retailing technology, Padala is passionate about making the forecourt safer. Since joining Allied, a world leader in forecourt technology and petroleum parts, Padala's primary focus has been directed towards two critical pursuits: decreasing gas pump fires and increasing payment systems security. In both efforts, he has used Artificial Intelligence (AI) and Machine Learning (ML) innovation to achieve notable results.
"Safety and security are paramount for the consumer experience," says Padala. "Convenience stores depend on their forecourts to drive loyal patrons in-store for shopping, whenever they stop to refuel. The more we can do to provide high-performance, seamless fueling experiences at the gas pump, the better we can drive profitability for our fuel retailing business customers and keep consumers safer."
Teaching AI to Detect Fires
Fire safety is a high priority for Padala, who was inspired by the accelerated emergence of AI tools to develop a new petroleum industry initiative known as "Fuel AI." He says, "We can't change the fact that gasoline is flammable. But we can now deploy intelligent automation to proactively monitor and mitigate risk in gas pump operations, and thereby reduce the incidence of fire."
Padala says that by integrating AI's predictive and responsive processes into fuel management systems, he is bringing fuel retailing safety technology into the 21st century and eliminating the potential for human error, which has historically been a drawback in traditional pump safety systems. "The first iteration of the manual shut-off valve in use today was introduced to gas pumps in 1939. Now, with today's advanced technology, we can minimize the fire risk associated with fuel stations' highly flammable inventories, and enhance operational efficiency by reducing the NPT (non-productive time) that results from fire accidents in the forecourt," he says.
The new Fuel AI innovation Padala designed automates a comprehensive monitoring and response system by integrating sensor fusion and real-time data processing with ML algorithms. First, a network of sensors, including vapor sensors, temperature gauges, flow meters, and flame detectors, are utilized in specific places throughout the fuel dispensing system to continuously gather data on critical parameters such as vapor levels, ambient temperatures, and fuel flow rates. Fuel AI then uses an advanced ML model, with algorithms trained in pattern recognition and anomaly detection, to analyze sensor data in real time and thereby identify potential fire risk factors.
Padala is particularly excited by the groundbreaking AI predictive modeling and risk assessment capabilities he has built into Fuel AI, which enables the system to leverage its historical data and predictive analytics capabilities to not only react to perceived threats but also to implement a proactive approach to risk prevention. For example, if Fuel AI detects a high-risk situation, such as a leak or equipment malfunction, it is programmed to take immediate and appropriate preventive action. Deploying an automated decision-making process governed by Padala's pre-defined rules, Fuel AI knows whether to automatically disable fuel pumps' active fire suppression systems or alert local fire departments and gas station personnel.
Teaching AI to make accurate decisions has been a highly challenging and training-intensive endeavor for the technology, as well as for Padala's engineering team at Allied. He says, "We have invested considerable time and expertise in using very diverse datasets so that the AI learns to make accurate predictions. This is especially important for the forecourt environment, where false predictions could generate life-threatening consequences."
Fuel AI is currently being implemented at select gas stations, while Padala and his team further hone the integration of the system with legacy station infrastructure. Because of the tremendous variation in the age and technology used by fuel retailing systems worldwide, he has designed Fuel AI as a modular and customizable component system to provide maximum interface flexibility. He is also working on enhancing Fuel AI to improve connectivity and scale by leveraging the capabilities of more advanced technologies such as IoT and cloud computing.
Protecting Payment Systems
While Padala is developing Fuel AI to provide for consumers' physical safety, he is also pioneering the use of AI with sophisticated monitoring technologies in gas station payment systems to protect customers' financial security better.
Payment card fraud in the forecourt is a $1 billion (USD) issue that impacts millions of consumers annually. In March 2021, LendingTree conducted a survey that found that 15% of Americans said they had been a victim of credit/debit card skimming at a gas pump. In the same survey, 43% of Americans said they had changed the way they pay for gas due to their fear of skimming, a crime in which criminals attach fraudulent card readers to payment terminals at gas stations to steal card data. Furthermore, the U.S. Federal Bureau of Investigation (FBI) has cited convenience stores and gas stations as hot spots for robberies (13.8%) and violent crimes (4.5%) in 2022.
Padala is leading initiatives to combat criminals at the pump in multiple proactive ways. First, to secure credit/debit card payment information, he has developed an innovative methodology that deploys robust, randomized encryption algorithms within in-memory buffers to ensure that sensitive payment data is only accessible to the active system occurrence when it is needed for the payment, and then immediately destroyed.
Additionally, Padala has developed a new Fuel Payment Application that is successfully speeding up secure and reliable payment processing for fuel retailers and their customers. An industry-first, this new high-performance, fuel payment solution gives retailers and point-of-sale (POS) providers a safe alternative to the time-consuming process associated with PCI SSF certification for software that processes card payments. Now, using Padala's innovation, retailers can go to market in just days instead of months.
Lastly, but equally important, is how Padala is leveraging cutting-edge AI technology known as Convolutional Neural Networks (CNNs)—an advanced form of computer vision—to continuously monitor and analyze forecourt environments and interpret data in real-time. With sophisticated training, CNNs built into fuel station security infrastructure can monitor fuel pumps to detect unauthorized access, tampering, or anomalies. By identifying suspicious activities and taking proactive measures, this new system can protect the forecourt infrastructure and the millions of staff and customers who populate fuel station forecourts.
Padala notes that while training AI and machine learning algorithms to identify and interpret fraudulent activity is a critical and significant challenge, it is just the first part of the puzzle; the second part is communicating that data to gas station attendants and/or appropriate local authorities. To do this, he is exploring unified, proactive monitoring systems that can integrate alert mechanisms, graphical user interfaces (GUIs), SMS notifications, emails, and/or other communication channels that will allow the system to instantly interpret data, generate alert notifications, and initiate pre-determined actions based on the perceived threat.
In these initiatives, Padala is at the forefront of fuel retailing innovation. "Safety and security at the gas pump, for both people and payment systems, is always my highest priority," says Padala. "Just as AI is transforming other sectors, new AI tools, such as predictive analytics, dynamic pricing, customer engagement platforms, fraud detection systems, and my new Fuel Payment Application are revolutionizing the fuel industry. We are now living in an era in which the legacy gas pump technologies will be replaced by automated, cloud-based solutions, with next-generation security such as facial and vehicle recognition programs. I am very proud to be leading my engineering team at Allied as we continue to pursue innovation and contribute to creating safer and more seamless fueling experiences."
