Cloud

5 ways AWS RoboMaker is transforming the robot programming game

Amazon's AWS RoboMaker can do a lot, but is it right for your robotics needs? Here are five ways RoboMaker is being used that might help you decide.

amazon-aws-deepracer.jpg

AWS DeepRacer

Image: Amazon

Amazon's new AWS RoboMaker puts a whole workshop of robotics software tools into the AWS cloud, making it easier than ever for businesses to automate their offices, workshops, and warehouses with robots. RoboMaker includes development tools, simulation environments, and deployment tools—all your organization has to do is supply the robots.

Knowing that RoboMaker exists doesn't answer one important question, though: Is it right for your business? These five use cases might give you a good idea of how you can make this innovative AWS robotics tool work for you.

SEE: Robotics in business: Everything humans need to know (free PDF) (TechRepublic)

1: Training ground vehicles and drones for use in 3D environments

One of the components of RoboMaker is a robust simulation suite that can virtualize 3D environments of various kinds. It comes pre-programmed with several models, including a retail store and a warehouse, but AWS customer Stanley Black & Decker used it for a different purpose: Training construction site robots.

"We are planning to use autonomous ground vehicles and drones to make the construction industry more productive while reducing construction rework costs," said Hamid Montazeri, Stanley Black & Decker's VP of SW engineering and robotics.

"Using a variety of imaging sensors, the collected data can be used to create 3D site models for planning and streamlining construction activities ... With AWS RoboMaker, we are able to easily test the robotics-related software applications in a cloud environment, and rapidly generate synthetic imaging data to train our 3D site model creation algorithms," Montazeri said.

In this instance, RoboMaker data collected in simulated environments was used to train machine learning models, thus reducing the time spent (and potential hardware damage incurred) during real-world testing.

SEE: AWS re:Invent 2018: A guide for tech and business pros (free PDF) (TechRepublic)

2: Eliminating onboard processing needs for internet-connected robots

Lea, an internet-connected smart walker built by Robot Care Systems (RCS) and designed for the elderly and disabled, uses a wide variety of sensors to keep its users connected to doctors, family, and emergency services. It also responds to voice commands, will speak back, and can navigate around obstacles.

Smart technology like that requires a lot of processing power, and RCS has been working with RoboMaker to shift the processing burden off of Lea and into the AWS cloud using RoboMaker's cloud extensions.

SEE: Must-read cloud computing coverage (TechRepublic Flipboard magazine)

RoboMaker can connect robotics hardware to AWS' cloud services like Kinesis for real-time data processing, Lex for speech recognition, and Polly for text-to-speech. Moving those services into the cloud gives RCS more liberty to design a sleek machine without heavy, loud, and energy-consuming hardware onboard.

3: Giving researchers instant feedback

One of the problems with controlling remote vehicles in space is that there tends to be a delay between sending and receiving a command. This can lead to confusion for operators, or in the case of a semi-autonomous rover, it could mean an accident that breaks multimillion dollar hardware.

NASA's Jet Propulsion Laboratory (JPL) has been using RoboMaker to build an open-source rover that it can get instant feedback from while testing. JPL has also used RoboMaker simulations to speed up production of new rover hardware by enabling it to instantly change hardware configurations, perform tests, and analyze data in the cloud.

Other researchers planning to use semi-autonomous robots of any kind could, if JPL's tests are any indication, use RoboMaker to speed up simulations prior to actual experiments, saving them time and resources.

SEE: Amazon Web Services: A cheat sheet (TechRepublic)

4: Finding particular objects in a 3D physical space

A breakout session at AWS re:Invent 2018 used RoboMaker to perform a simple experiment: Train a wheeled robot to find and identify a particular object—in this case a dog.

The experiment itself may have been basic, but it illustrates another use case for RoboMaker: Training robot software to perform complex tasks like object recognition in tandem with navigating a 3D space.

SEE: AWS RoboMaker: A cheat sheet (TechRepublic)

If the session presenters were to continue training their wheeled dog-identifying robot, it could be used to find products in a store without the need to develop a full floor plan, find people lost in the wilderness, or aid medical professionals in performing examinations in a variety of environments and conditions. Those various robots could also be trained to do all of that with a high degree of accuracy before ever leaving the virtual world.

5: Giving educators a machine learning tool

Amazon revealed a new hardware product at AWS re:Invent: DeepRacer. It's a 1:18 scale race car that uses reinforcement learning to teach it to navigate race tracks without the need for a driver.

A DeepRacer can be pitted against other DeepRacers to see whose machine learning algorithm is the best—Amazon even plans to have an AWS DeepRacer league starting in 2019. That said, those machines need training, which they can get in a RoboMaker simulator custom built just for DeepRacers.

SEE: What is machine learning? Everything you need to know (ZDNet)

For only $249 on Amazon, DeepRacer is a great tool for students and independent learners interested in the future of machine learning, and it highlights another great feature of RoboMaker: It's creating a more accessible robotics programming environment.

RoboMaker puts robotics programming, simulating, and managing robots in reach for anyone with an AWS account, whereas before it would require weeks or even months of software development and hardware engineering to bring an organization to the point of even beginning to build code for a robot.

AWS RoboMaker is available now, and anyone with an AWS account (even free-tier customers) can try it out today.

Also see

About Brandon Vigliarolo

Brandon writes about apps and software for TechRepublic. He's an award-winning feature writer who previously worked as an IT professional and served as an MP in the US Army.

Editor's Picks

Free Newsletters, In your Inbox