In the robotics space, filled with options that are up to performing almost any task, autonomous mobile robots (AMRs) stand apart due to their abilities to negotiate their environments and understand the context in which they’re working without a controller to direct and oversee them. Through a network of software, middleware, devices, and artificial intelligence (AI) including machine vision, AMRs have become valuable assets in businesses, such as:
There are numerous ways AMRs deliver a return on investment (ROI) in these and other use cases – often substantially and rapidly. Unlike fixed robots, AMRs can perform tasks wherever they’re needed, whether that’s across the shop floor or in a different facility. They can work with employees or teams to
carry materials or perform routine tasks, saving human workers steps and time so that they can be more productive. AMRs may be multifunctional – it’s not necessary to design them to perform only one task – so they don’t have to sit idle. They can be trained and deployed within one to two months, and they’re scalable – you can deploy one, then add additional AMRs as needed.
Considering the benefits and return AMRs deliver in a range of use cases, expect to see adoption take off over the next five years. ResearchAndMarkets.com estimates the global autonomous mobile robot market will grow to $145.5 billion by 2026, representing an impressive CAGR of 24.6%.
If your business has decided to be a part of that growth trend, it’s important to acknowledge that successful AMR projects depend on four fundamental areas:
1. Robot Operating System (ROS)
Nearly one million, or 55% of all robots sold in 2024 will be using ROS. ROS which is neither a robot nor an operating system, is a set of tools and frameworks for developing robotics applications. ROS is one of the most crucial parts of an AMR project. Open Robotics’ ROS is a collection of tools, libraries, and conventions that allow robot developers to leverage existing work to accelerate their project timelines. ROS 1 originally didn’t provide developers everything needed for industrial robot projects, those that require mission-critical security, integration with other robots and operational technology, and scalability. That demand drove the development of ROS 2, the choice for industrial AMR projects.
2. Open-source Technology
Proprietary technologies can put the brakes on achieving your vision for full AMR functionality and the robot’s ability to work with other systems. Open-source technology enables a fully connected ecosystem capable of the free flow of data to and from robots, industrial equipment, systems, the cloud, applications, and people.
Eclipse Cyclone DDS (Data Distribution Service), an open-source Eclipse IoT incubator project, for instance is advancing open-source robot development. We are proud to be a large contributor to Eclipse Cyclone DDS which is designated as a tier-one ROS 2 middleware by the ROS 2 Technical Steering Committee and also the default for the ROS 2 Galatic Geochelone release. Open-source projects like Eclipse Cyclone DDS, and even Zenoh, support ROS 2 and help make AMR development faster, easier, more secure, and more reliable.
3. Edge Computing
AMRs must process high volumes of data from multiple inputs to perform accurately and reliably. Inputs can include AI vision – object detection, ML training and 3D sensing, LiDAR and sensors, communication with other robots, PLCs and industrial systems and much more. AMRs must make split-second decisions, continuously computing in real-time to make decisions and safely interact with humans. For example, when an employee steps into the path of a warehouse AMR that transports materials or when a hospital AMR detects life-threatening patient vital signs, transmitting data to and from the cloud with an uninterrupted connection isn’t practical when a real-time response is necessary.
A strategy of both edge and cloud computing is the better solution. Developers can take advantage of the cloud to manage the large data sets needed to train AMRs. Then, deployed robots can operate at the edge, leveraging the autonomous computing power needed to work instantly. Our ROScube ROS 2 Robotics Controllers for instance combines ADLINK’s leading military-grade DDS software with rugged edge hardware that sits directly at the edge of the AMR, for AMR navigation.
4. Partner Ecosystem
Successful AMRs are most often the result of building the right partner ecosystem, leveraging many different technologies. Embarking on an AMR project alone – if it’s even possible for your internal resources to do so – would take much more time than necessary. It would also mean that you’d undergo trial and error that could be avoided by leveraging experienced partners’ expertise to help keep the project on track. The right partnerships enable you to deploy an AMR more quickly and help ensure that it works reliably and accurately in real-world situations, and gives you the results and the ROI you need.
A cool example of a partner ecosystem working together for AMRs is the Indy Autonomous Challenge, the world’s first high-speed, head-to-head autonomous race coming up at the Indianapolis Motor Speedway (of course we’re proud to be the official edge computing sponsor!). And yes, you read that correctly. This is the first-ever race fully powered by computer, no humans are racing, they’re instead engineering and developing.
The racecars are essentially really fast robots. Every racecar is a modified Dallara IL, retrofitted with hardware and controls to enable automation. An ecosystem of open-source development contributors are able to make it happen including both Open Robotics and the Eclipse Foundation of course, but also Apex.AI, The Autoware Foundation, Industrial Technology Research Institute, Open ADx working group and more.