Amazon checks its own fleet with AI gateways • Bestelauto.nl

Amazon has launched new artificial intelligence (AI)-based technology that can detect even the smallest irregularities in Amazon delivery vehicles – from deformed tires and undercarriage wear to bent or warped body panels – before they become a problem on the road. New Automated Vehicle Inspection (AVI) technology will help fleet managers who previously relied solely on the human eye and manual inspections for day-to-day safety. Amazon launched AVI technology in partnership with technology startup UVeye in the US, Canada, Germany and the UK.

“The last thing I want is for something preventable to happen, like a tire blowing out because we missed a defect that wasn’t visible during the morning inspection,” said Bennett Hart, an Amazon Delivery Services Partner (DSP) who owns logistics company Hart Road. “This technology increases the safety of our fleet.”

Improved safety for the nearly 280,000 drivers who deliver packages to Amazon customers through Delivery Service Partners like Hart Road is just one of the benefits of this technology. Another advantage is that AVI benefits can be scaled up, which is useful because DSP drivers deliver 20 million packets to customers every day.

“A compelling benefit of AVI is the suite of insights the technology provides fleet managers,” said Tom Chempananical, global fleet director at Amazon Logistics. “It can track detected vehicle problems and see if they occur repeatedly on a particular route.”

How AVI works

At the end of each work day, the DSP driver drives through an AVI arc and a series of boards equipped with sensors and cameras. “When you go to the doctor, you expect to see the results of a scan; we do more or less the same thing but for vehicles,” said Amir Hever, CEO of UVeye. When the vehicle is traveling at a speed of 5 km/h, the AI ​​system performs a thorough scan of the vehicle within a few seconds, identifies problems, classifies them by severity, and immediately sends the results to the computer. From there, the DSP can determine what repairs and services they need to perform so that a well-maintained vehicle can be back in service the next day.

While this technology was originally created to scan the undercarriage of vehicles at borders and security checkpoints, it now uses AI to look for more specific and minute details, such as vehicle damage. AVI relies on machine “stereo vision,” which means it uses two points of view to create a full 3D image, and deep learning, a subset of machine learning in which layered neural networks mimic the learning process of the human brain.

“We can’t just pull out this UVeye solution and start using it,” Chempananical said. “Considering the unique demands of Amazon’s massive fleet of more than 100,000 vans, ranging from specialty vans to Rivian electric vans, we worked directly with UVeye to train AI models and algorithms according to the strict standards of Amazon’s Roadworthy Guidelines to keep the wheels rolling. turn safely.”

A game changer for fleet maintenance

By carrying out inspections more quickly, accurately, systematically and objectively, AVI has discovered hidden damage patterns. For example, 35 percent of all problems are caused by tires. These problems include cracks in the sidewalls, chips and nails in the tread, problems that previously were not easily visible through manual inspection. With AVI, DSP receives notifications to replace tires before they become a bigger problem. Preventing blown or flat tires on the road will improve safety and the delivery experience for drivers, ultimately eliminating potential delays for Amazon customers.

“The great thing about AI is that every crash is then fed back to an API that can train the model and improve detection accuracy. So the more AVIs used, the better it will be,” said Chempananical.

Analysis is performed through Amazon Web Services (AWS), where terabytes of image and vehicle data are processed and stored using Amazon Simple Storage Service (Amazon S3), AWS Lambda, Amazon Elastic Compute Cloud (Amazon EC2), and other services. The output is processed and shared via API in less than a minute. Amazon has integrated the output into existing DSP experiences using Amazon Fleet Portal and shows fleet managers the detected issues along with photos and suggestions for improvement.

Learning everything there is to know about Amazon fleet tires is just the beginning. Many other details are examined with this digital microscope, including identifying potential trip and fall hazards on load ladders and finding damage to hazard lights.

“This technology could become a new standard for vehicle inspection,” Hever said. “Amazon is always thinking about scale and the fact that we are now working with one of the largest fleets in the world helps us grow. This partnership is a reminder of what we have built and how we can bring it to life at hundreds of delivery stations, dealers and other locations.”

What’s the next step?

In the future, this technology could be used to support all types of delivery vehicles, from bicycles to drones, and could inspect not only the outside but also the inside of the vehicle.

“The potential for implementing WIP is very large,” said Chempananical. “This could also impact our routing and geolocation technology if we see repeated vehicle failures in the same location.”

With AVI as a key part of his DSP toolkit, Hart is confident he can run his vehicles and spend less time on inspection tasks, which can now be automated. This technology has provided a way to grow their business and reduce fleet wear and tear in the long term, as well as providing peace of mind for entrepreneurs.

“There is a lot that goes into running a delivery business and properly maintaining vehicles, and manual inspections are not 100 percent accurate or objective,” Hart said. “With AVI, I know I can spend more time with my drivers and focus on creating the best delivery experience for them.”

Rebecca Burke

"Coffee trailblazer. Analyst. General music geek. Bacon maven. Devoted organizer. Incurable internet ninja. Entrepreneur."

Leave a Reply

Your email address will not be published. Required fields are marked *