30 Jun 22

Fleet Edge helps Amazon avoid congestion and improve driver safety

After two years of development and piloting, Amazon has introduced Fleet Edge for drivers and Delivery Service Partners (DSPs). The core aim is to increase driver safety and route around roadworks, congestion and avoid accidents and downtime. 

In the past, it would take weeks for Amazon's routing software to be updated for any change on the road. The long update times were challenging for drivers as re-routing was causing navigational problems. 

Having invested more than $1 billion in new technologies to accelerate its DSP program, Amazon is now determined to eliminate all navigational challenges through real-time routing while increasing driver safety with Fleet Edge. 

How does the new technology work? 

Fleet Edge automatically detects changes in roads, traffic signs and highways; mapping software continuously updates the on-road information, enabling drivers to take advantage of the most recent routing data during deliveries. 

Fleet Edge combines in-vehicle computers, street-view cameras and GPS receivers to constantly track and update changes occurring on the road, minimising downtime and incidents in fleet operations. 

When a driver is on the road, the Fleet Edge system automatically detects any changes on the route, including newly constructed roads, traffic signs, closures and buildings. The map of a particular area is updated accordingly and expanded to offer better navigation for future routes. The system works efficiently, as Amazon says Fleet Edge has added more than 33,000 new signs to the mapping system in the last few months. 

DSPs regard the Fleet Edge technology as a game-changer thanks to the many advantages it offers: 

  • Drivers can avoid risky manoeuvres, 
  • The need for U-turns significantly decreases, 
  • Bypass road hazard. 

Fleet Edge also remarkably assists navigation by announcing impending turns sooner, as GPS location accuracy has increased by 260%. 

Further development through machine learning

Amazon doesn't disregard the privacy of customers and drivers, as Fleet Edge automatically removes or blurs all personally identifiable information (PII). This includes any individuals who walk in front of vehicles. 

The system combines artificial intelligence (AI) and machine learning (ML) to scan roads before the updates are added to maps and the data gathered is encrypted. 

Dedicated to improving the experience for customers and DPSs, Amazon Last Mile, responsible for developing planning software for the delivery fleet, launched the Amazon Last Mile Routing Research competition with the Massachusetts Institute of Technology (MIT). The program aims to train machine learning to predict delivery routes using Amazon's historical data of deliveries and information regarding more than 4,000 routes set by drivers. 

Image: Shutterstock

Authored by: Mufit Yilmaz Gokmen