Digital Twins Paving the Way for Autonomous Last Mile Delivery

Amazon is testing delivery robots in real time–and with a digital twin.

Scout is a delivery robot that Amazon is testing in a Seattle neighborhood, but before physical Scout takes to the streets, a digital Scout has already rolled through the digital version of the neighborhood thousands of times in different scenarios. This allows Amazon to test without severely disrupting the neighborhood–and also lowering the risk for Scout.

Digital twinning is not new, and many supply chain companies are starting to implement digital twins for testing and long-distance maintenance.

This example from MHI Solutions magazine highlights the power of predictive analytics, big data, and digital twinning. “Schindler Elevator Corporation is having more success with IoT sensors and virtual reality. It installs sensors in every elevator and builds a digital replica of it from those sensors. Using the digital twin, Schindler technicians can detect when something is about to fail and make a proactive service call equipped with the right parts. “Since you have advance notice that something is going to fail, you have the opportunity to make sure that you have the necessary parts moving through your supply chain,”[says Randy Bradley, Assistant Professor of Information Systems and Supply Chain Management at University of Tennessee].

In our 2019 Annual Industry Report, Elevating Supply Chain Digital Consciousness, we take a look at the different stages of digital adoption and how that affects supply chains (you can read the full report at and created an assessment tool for companies to determine their own Supply Chain Digital Consciousness Index to know where they are in their digital journey.

Digital requires a lot of data, and big data is essential in building a digital twin. Amazon has built an eerily accurate version of the Seattle neighborhood using a combination of scans of the neighborhood and extrapolating data from its business and cloud computing platforms according to Wired.

Right now, any potential robotic delivery programs are still in infancy, facing scalability, reliability, and autonomy issues. But even these test pilots are starting to affect the real world. Not long after Amazon announced that testing was beginning, Seattle’s governor signed a bill into law that would regulate the behavior of future delivery bots.

Data and machine learning are also coming into play when Scout is about and about live. Gathering as much data as possible is key to making Scout self-reliably and eventually detaching from a human driver. Scout will need to gather data in training through the digital twin and also gather data live to be able to avoid pedestrians and navigate new obstacles, and this data gathering often requires repetitive tries of the same action, which can be annoying if it’s happening around non-testing humans. The digital twin allows Scout to train problems (like mounting curb ramps) without disturbing the physical neighborhood.

Digital is here. It’s still in early adoption, but it’s not going away. The question that remains to be asked is if your company will be able to keep up. That’s why our Annual Conference is centered around the Annual Report, so that we can help you find your entry into digital if you haven’t started, or discover new ways to keep elevating your supply chain digital consciousness. Learn more and register at