Robots, be it softwarerobots (RPA) or physical robots, do not function without an adequate infrastructure. One of the largest infrastructure providers in the world powering robotics is Amazon Web Services (AWS). Rocking Robots had the opportunity to talk to Eric Anderson, general Manager AWS Robotics and Autonomous Services at AWS about their strategy supporting robotics.
What role does AWS play in the process of automation?
“AWS Robotics includes a broad and deep set of service to build, test, manage and deploy autonomous robots. High performance autonomous robots are equipped with software for capabilities like navigation, perception, data processing, analytics, and, machine learning. This software requires expertise to build and must be comprehensively tested. Once the software is deployed, robot managers needs to ability to remotely manage, operate, and update robotics to maintain and optimize functionality. AWS Robotics provides robot builders tools to enhance robot functionality, test robots in simulation, and to securely deploy, connect and monitor fleets of robots.”
How is AWS supporting robotics developments with their infrastructure?
“We believe AWS can make robots ubiquitous by providing a portfolio of cloud-enabled solutions that enable failure resilient, easier to operate, and smarter robots. Cloud-based simulation is essential to accelerate, scale, and improve robotics development. Real world testing is not robust, cannot be easily scaled, and is expensive. Robotics companies not using simulation, will be limited in their efforts to advance their robots development.”
“AWS has invested in AWS RoboMaker to make it easy for developers to adopt and use simulation without managing any infrastructure, enabling them to run hundreds of tests with a virtual robot, design robot motion and control behavior, and train machine learning models for perception. These simulation-enabled capabilities unlock the ability for robotics companies and AWS customers like iRobot, Roomie, JetBrain and Milvus to build high quality, autonomous robots faster and with fewer resources.”
“AWS investments in edge computing and AI/ML have have resulted in dozens of AWS services are transforming how robotics companies (robot builders) and their customers (robot operators / users) leverage the cloud today. These services are enabling companies to build intelligent, autonomous robots like Ghost Robotics, Amazon Robotics, and Woodside with integrations to AWS services like Amazon SageMaker, AWS Wavelength, and Amazon Kinesis.”
“AWS IoT provides reliable, secure, and scalable tools to deploy, connect, and operate robots at the edge. Customers examples include Seafloor Systems Inc. and Roomie. Securely connecting to robots, then updating software remotely, is a time-consuming and error-prone problem to solve that often prevents robotics companies from leveraging the full potential of the cloud. With AWS IoT it is easy to set up a secure and reliable connection and update robot software over-the-air. With AWS IoT Greengrass 2.0, developers can extend cloud services onto robots so they can act locally on the data they generate, while still taking advantage of the cloud based solutions like fleet monitoring and dashboards with AWS IoT FleetHub, video streaming & analysis with Amazon Kinesis Data Streams, and data ingestion & analysis with Amazon Kinesis Data streams.”
What trends do you see with your customers?
“Robotics customers are seeing tremendous demand for solutions from both traditional adopters of automation, like fulfillment and manufacturing, but also from new sectors including transportation, natural resources, and agriculture. To deliver, our robotics customers are looking for solutions to make their equipment more adaptable and enable them to scale. Just a few years ago, if you had a couple dozen robots you were an advanced adopter of robots. Today, I regularly talk with customers with hundreds of robots and they are looking to deploy thousands. This is where AWS helps. Whether it’s AWS RoboMaker for simulating new robots or algorithms, AWS IoT and Greengrass for scaled data collection and device management, or AWS SageMaker for machine learning, AWS’s unique portfolio of technologies simplify development of robotics and accelerate the delivery of automation solutions.”
What developments do you expect in robotics in the upcoming years?
“Amazon is a big contributor-to and user of ROS. We made a number of contributions to ROS2 and see that ecosystem as an important part of how robotics development can be accelerated. I think ROS2 has a bright future. Interoperability between robots and industrial systems is becoming a bigger deal. We’re seeing more hybrid fleets – AGVs (automatic guided vehicles) and AMRs (Autonomous mobile robot) running the same workloads — and robot-to-robot process integrations. When you look at how these systems are built they can get really complex and highlight interoperability opportunities between different types of equipment and core manufacturing systems. It will be really interesting to see how these new architectures evolve and what role hybrid edge-cloud services play”
“We also seeing this interesting trend in adaptable robotics. It used to be that when you deployed a robotics system it did one thing and required significant reprogramming or hardware changes to adapt it to something else. Today, with the diversity of sensors, and the ease with which you can deploy ML-based perception systems in AWS, we’re seeing equipment be adapted to new use cases faster than ever before. Amazon’s fulfillment organization recently showcased an example of a robotic arm adapted with sensors and custom ML-models to manipulate packages.”
“The last thing I’d highlight is the continued importance of robotic simulation throughout the development lifecycle — from early concepting, to deployment, to optimization at scale. Robotics is hard because you have to deal with the real world — actual physics! Simulation is this great accelerator that let’s you innovate at the pace of software development. Robotics organizations who aren’t adopting simulation in a big way are missing out one of the biggest ways to help their teams deliver better product faster”
In the area of software robots and end-to-end automation AWS partners with UiPath, to deliver enterprise-ready cloud infrastructure, AI services, and cloud contact center solution. UiPath is an AWS Advanced Technology Partner in the AWS Partner Network. For example, UiPath integration with Amazon Connect enables a zero-touch customer experience, so you can reduce call volume and free up agents to work on complex customer issues.