Embedded Insiders Podcast: AMRs, AI and edge computing

By 2023 it’s expected that eCommerce will drive two-thirds of retail growth, up from 50% in 2019. This tremendous growth is putting a great amount of pressure on operations – from production, to fulfillment, to distribution, and all eyes are on autonomous mobile robots to help enhance productivity and efficiency at each stage. 

But how does one even get started building autonomous robots to help automate logistics?  

From Automation to Autonomy 

That’s where Industry 5.0 comes into play. Whereas industry 4.0 paved the way for enhanced “smart” automation, the follow-on specification takes it to the next level by placing autonomous mobile robots (AMRs) into the same working environment as humans.  

AMRs can assist with the ever-changing environment and dynamic workflows of a warehouse and other industrial settings. In fact, in a Gartner survey of retail CIOs, 87% indicated robotics for warehouse picking as the number 1 use case. With AMRs, manufacturers and distributers can expect a reduction in operating expenses.  

“Integrating IoT Edge computing and AI is a key part of this design,” said the director of product management for autonomous machines at NVIDIA, Amit Goel. “We are building a platform that brings intelligence to these autonomous systems. Distributing this intelligence across the different components of your warehouse makes everything work synergistically, which isn’t possible with a single compute infrastructure.” 

Having AMRs and humans collaborating safely in an environment requires the edge – both hardware and software for real-time, failure-isn’t-an-option computing. 

Edge Hardware and Edge Software Working Cohesively 

We can’t expect the highest level of efficiency without hardware and software coinciding. NVIDIA is currently in the process of developing application-specific software development kits (SDKs) which will allow developers to create application, or industry-specific AMR solutions. The integration of these solutions will be streamlined because they will all operate on the same core GPU-based platform. 

As you would expect, there are a wide range of needs within a manufacturing plant or warehouse distribution center. When it comes to automation, the needs could be as simple as using machine vision to scan products to ensure organization, or in a more complex environment, using data acquisition and machine vision within the AMR to travel throughout the warehouse for optical inspections, coming in contact with other AMRs and humans along the way. 

We’re at the forefront of the AMR space with leading edge computing solutions designed specifically to build and deploy AMRs safely in industrial environments – from different core options for heterogeneous computing, to different modules and even integrated edge platforms, we have you covered. 

Organizing Your Data in Real Time 

We’re seeing logistics systems create mass amounts of operational data, and this data needs to be processed in real-time, at the edge, on, near or around the data-producing “thing”. All of the equipment communicates with the use of specialized middleware, such as Eclipse Cyclone DDS, making sure point-to-point communications is successful within a given time.  

Check out the second episode of the Embedded Insiders Podcast where we discuss more on edge AI and AMRs and stay tuned for episode 3! 

To learn more about ADLINK and NVIDIA GPU solutions, visit our website

Author: Zane Tsai
Author: Zane Tsai

Director of Platform Product Center, Embedded Platforms & Modules, ADLINK Technology