A collaborative mobile robot from Robust.AI. (Source: Robust.AI.)

In response to global supply chain challenges and workforce shortages, as well as a profound change in the way people buy goods due to the COVID-19 pandemic, companies are increasingly relying on robots to do the work needed to meet customer demand. This means that more workers will be placed in the pathways of more robots more often. But what robots are best suited for working alongside human workers? Common robots currently found in warehouses such as automatic guided vehicles (AGVs) and automated mobile robots (AMRs) have proven to be useful, but they come with limitations. A better solution is collaborative mobile robots (CMRs).

Robots are found in workplaces from the warehouse floor to the hospital hallway. But as workers are being asked to share space with these robots, they are also placing more demands on the technology. Workers don’t want machines that will only perform simple tasks and are more of a nuisance than a help: they want to work with the robots, not around them.

Bears and Goats

While useful for specific tasks, AGVs and AMRs have some significant challenges to overcome.

AGVs have specific functions and perform their tasks along strictly prescribed and preset routes, usually laid out along wires or magnets embedded in the floor. It’s as if they follow imaginary train tracks that they cannot deviate from. They have simple sensors that enable them to avoid hitting obstacles in their path, but they aren’t sophisticated enough to navigate around them—they just stop and wait for the obstacle to be removed.

In his keynote at the 2022 Digital Factory conference, robotics expert Rodney Brooks described AGVs as bears—you want to make sure a fence or a cage is between you and the robot or you run the risk of getting injured. There is very little, if any, human-robot interaction.

AMRs are more sophisticated. Early models were essentially collaborative robot (cobot) arms mounted on wheels, but they’ve grown in complexity. These robots are a simple, efficient and cost-efficient way to automate the way materials are handled and transported within a facility such as a warehouse—basically replacing the need for a human worker to load and push a cart around the facility.

AMRs use sensors and software to scan and interpret their environment, which enables the robots to navigate efficiently, working their way around stationary and moving obstacles. They can work well with operators on specific tasks such as picking and sorting objects. But because they are expected to inhabit the same space as human workers, they must meet high responsiveness and safety protocols.

Brooks likens these robots to goats: they do their own thing and pay little attention to the humans in the room. An AMR likely won’t hurt a human, but the human is nothing more than a point-in-cloud obstacle, or an object in their LiDAR, that is to be avoided or worked around.

Where AGVs and AMRs really come up short is in their interactions with humans. They function either in entirely separate workspaces to prevent them from injuring workers or are limited to specific jobs where any teamwork with humans is constrained to specialized movements and operations.

As a result, the humans and robots largely work separately from each other. Even AMRs, which do share workspaces with humans, can frustrate workers if they get in the way, becoming more of a hindrance than a help.

Human workers need to feel comfortable around robots and confident that they will perform their work effectively and safely. When robots ignore social cues that humans take for granted, it’s not only a matter of discomfort but also productivity. If a robot is constantly slowing down or stopping to negotiate an obstacle, it can introduce snags into the human’s workflow.

Collaborative Mobile Robots: An Evolution of Human-Robot Interaction

The next big challenge in humans and robots sharing the workspace is for them to be able to actually work together. This will, of course, require a far more sophisticated and intuitive robot that is able to read human body language, identify and respond to social cues and anticipate what their human coworkers will do next.

That’s what the CMR is intended to be: a mobile robot that can read human body language to identify what workers around it are doing and respond appropriately. These products often deploy technologies such as LiDAR sensors, 3D depth cameras, and other sensors, and use artificial intelligence (AI) to get a clear view of the space around them and map out for themselves the best and most efficient way to carry out their tasks—while taking cues from humans about how to behave.

“The analogy here is a service dog,” said Brooks, whose company Robust.AI is developing its own CMR as well as a software platform to support it. “It obeys you; you can modify its behavior and it’s there to help you.” Rather than treat human workers as obstacles to be overcome, the CMR treats them like collaborators … or dance partners.

(Source: Robust.AI.)

In his keynote presentation, Brooks cited an example of a robot encountering two people who are facing each other and socially distancing. The software his company has developed enables the robot to interpret that the humans are interacting with each other, so instead of taking the direct route between them (and disrupting their conversation), it moves around them. Had the two people been facing away from each other, the robot would have taken the straight path between them.

Five Killer Features of CMRs

CMRs have five distinct advantages for companies aiming to bring automation into their facilities.

Intelligence

Thanks to deep learning and AI, CMRs generate and process information to perform tasks efficiently with little or no need for a human operator to tell them what to do. These machines can also self-manage and make decisions on the fly. As a result, the robots can execute repetitive or hazardous tasks and adapt their movements to the humans around them in real time.

Connectivity

An increase in processing power needed for intelligent automation also means that CMRs can use machine-to-machine communication to interact with humans through multiple and integrated interfaces such as touchscreens and sensors. In contrast, many AMRs do not have this level of functionality or interactivity.

Flexibility

CMRs are also quite adaptable, adjusting the way they work depending on changes to their work environment. For example, a CMR could adapt as the layout, organization and needs of a warehouse change in response to fluctuating market demands.

Improved Safety

Intelligent CMRs that can handle and process massive datasets in real time are better able to predict human actions and perform high-precision movements in response to humans—making these robots safer for people to work alongside. These machines also have highly responsive and reliable failsafe mechanisms.

Easy Deployment

Should a business want to introduce AMRs in its facilities, it will need to bring in a team of professionals to set up, configure and control the robots. In contrast, CMR deployment can be done by anyone without the need for technical training. Rodney found success in this approach with iRobot, another one of his companies that produces the Roomba floor cleaner. The Roomba doesn’t need a systems integrator to install it in a home—the resident who purchased the robot can merely plug it in and press a button to put it to work.

The future of robotics is clearly heading toward collaboration—a world in which humans and robots work side by side rather than on opposing sides of an invisible bear cage or with the disregard of a grass-gnawing goat. Soon, perhaps the animal we’ll associate most with robots will be the stalwart dog, man’s best friend—for now.