TankU Technology aspires to become the industry standard for road side transactions and services
Rapid improvements in artificial intelligence as well the abundance of sensors in various environments, serve as a fertile ground for increasing value for customers and businesses alike, as well as discovering new and lucrative commerce options. Another important trend already disrupting conventional business models is the emergence of the autonomous and smart vehicle, and the innovative business models surrounding them. In this world of smart technology and autonomous cars, a world of immediacy and higher demands for security and control, TankU aspires to take the lead and become the industry standard for road side transactions and services.
The TankU system combines data from image sensors, mobile sensors in smartphones (in the future also sensors in vehicles) and from hardware already present in gas stations. The system enables automatic labeling of video data in unique environments and scenarios. The system employs cutting-edge algorithms in machine learning, deep learning and computer vision to recognize vehicle plate numbers, vehicle features and customer features, achieving detection accuracy dozens of percent higher than existing solutions and therefore enabling the use of low quality cameras. The system’s learning enables improved identification accuracy in video, and as time progresses also in environments without multiple sensors; in highly variable conditions of camera positions, angles and dirt; for localization and navigation with mobile sensors.
The company has submitted three patents to protect the technology, in addition to the implementation of our innovative service.
TankU uses this technology to bring novelty and innovation to the gas station domain. Today, adding an extra security component to fueling systems is expensive and cumbersome, whether it is a vehicle-installed RFID device or a long and arduous process for identification via a fueling mobile app. TankU’s unique technology adds this security element, identifying vehicle and driver. Such an element increases the level of security, and at the same time significantly improves user experience by comparison with other payment methods: credit card, fuel card, mobile apps, etc.
TankU aims to set the standard for purchases performed via mobile sensors (e.g. smartphones). The TankU solution combines data from mobile sensors and cameras. The low cost of camera facilitates and their incorporation in any point of sale, allow the acquisition of unique data for each transaction. Transactions performed using the TankU solution are both the most secure and deliver the best user experience. The process developed by TankU allows ‘one tap’ refueling via a mobile app, activating the fuel pump as soon as the customer exits the vehicle, cutting up to 95% percent in activation time.
The industry standard for visual identification systems is deduction, as existing identification systems are developed and trained using existing databases of labeled images and/or video. Domain-specific databases are expensive and hard to acquire, and non-specific ones contain data that is diverse and not necessarily relevant to the task at hand, leading to mediocre results. The sensor fusion employed in the TankU system enables the continuous labeling of data for each individual sensor (cameras, mobile sensors, etc.), and thus enables continuous training and improvement. This feature allows specific learning for each individual sensor, as well as induction and deduction. This unique learning process allows for fast learning with detection accuracy dozens of percent higher than existing solutions in the environments and scenarios tackled by TankU, even when using irregular sensors. This algorithm can later be applied to achieve high detection accuracy even in environments containing cameras and no other sensors.
