The future of delivery will include artificial intelligence. This is the vision of the e-commerce giant which will deploy 1,000 electric delivery trucks equipped with a computer-aided parcel retrieval system by the beginning of next year. “If searching for packages in the back of trucks takes only a few minutes at each stop for a delivery person, that time can double when it comes to delivering packages to Amazon’s more than a hundred customers every day.” Amazon notes.
Accordingly, the company has developed an AI-powered solution called Vision-Assisted Packet Retrieval (VAPR). However, delivery drivers will no longer have to spend time organizing packages by stopping, reading labels, or manually checking key identifiers such as the customer’s name or address to ensure they are getting the correct packages. All they have to do is look for the green light from the VAPR, grab it and go.
Easy tagging of delivery drivers
The system has been simplified as much as possible to assist the delivery person: parcels are marked with either a red “X” or a green “O” thanks to the VAPR system. Using an audible and visual signal, the VAPR system prompts the driver to ensure they have found the correct packages, before entering the loading zone. Using this technology, drivers no longer need to use a mobile device throughout the process.
Behind this solution is a team from Amazon Transportation that took the initiative in 2020 to imagine a way to combine technology with the daily lives of drivers. VAPR uses Amazon Robotic Identification (AR-ID), a form of computer vision originally developed for fulfillment centers, to automatically identify items during stocking or picking operations.
AR-ID replaces manual barcode scanning. Techno can locate multiple barcodes and decode them in real time. Therefore, the VAPR system was designed by training ML models to recognize different labels and packages, under different lighting conditions and package characteristics.
Adapting to the environment
“We had to consider factors specific to the delivery experience, such as lighting and space constraints inside the trucks.”“, explains John Colucci, Product Manager, Amazon Transportation. In fact, the problem of optimizing this technology inside the truck quickly emerged.
This technology is therefore combined with projectors and cameras designed specifically for cars, all of which are integrated into the truck’s delivery route navigation system. VAPR uses several AWS technologies on the go, including the SageMaker platform and Greengrass IoT. Four years later, and after multiple tests with drivers, the VAPR will finally be deployed in 1,000 Amazon electric delivery trucks, manufactured by Rivian.
Significant time savings across the entire fleet
Based on initial testing, Amazon says it has seen a 67% reduction in physical and mental effort seen by drivers and a time savings of more than 30 minutes per trip. “Before, it would take me two to five minutes to unpack a box and organize packages for the next stops. Now, with VAPR, it only takes me one minute.” explains Bobby Garcia, a driver with Bloomfield Logistics, who tested the solution in the North Boston, Massachusetts, area.
With more than 390,000 drivers associated with its Delivery Service Partner Program worldwide and more than 100,000 vans in its fleet delivering millions of packages daily, Amazon relies heavily on its VAPR system in terms of time savings.
Selected for you
