« Back To Industry Insights
Association for Advancing Automation Logo

Member Since 1974

LEARN MORE

Content Filed Under:

Industry:
N/A

Application:
N/A

Automation’s Biggest Challenges: Full Human-Robot Collaboration

POSTED 09/04/2024

The rise of collaborative robots (cobots) has made human-robot collaboration (HRC) an everyday occurrence in manufacturing, warehouse, and logistics facilities.

But human-robot collaboration takes many forms (see sidebar), including caged cobots that receive rare visits from humans or simple human-robot coexistence in a shared workspace.  Robots and humans can even work on the same item in the same workspace. 

At the top end of the scale is Level 5 HRC in which robots and humans can safely work on the same part at the same time, with the robot dynamically responding to human actions.

Definitions of HRC vary slightly, but in practice most collaborative industrial applications involve sequential (or Level 3) collaboration. A human performs a task — such as feeding parts to the cobot in a shared workspace — and the cobot performs its role, such as placing those parts into a CNC machine, turning it on, and placing the finished parts in a bin.

However, as cobots get smarter, boosted by AI hardware and software, including foundational robot platforms, the possibilities for Level 5 HRC seem almost limitless.

Some futuristic scenarios even envisage robots monitoring human stress levels and emotional states in real-time — possibly via EEG sensors. They would then use this information to control the robot’s speed to ensure that the application is as ergonomic as possible for human workers.

Image courtesy of SICKSafety first

It’s easy to get caught up in semantics when defining what “full” HRC entails but the parameters of space and time are fundamental to understanding different levels of HRC, says Jayson Pestow, a robotics business consultant at SICK, a leading manufacturer of sensor and safety solutions for industrial applications.

“In ‘Level 5’ HRC, there’s a common space and time where the human is present and the robot is active and moving. In this dynamic environment, the robot responds to where the human is, what the human is doing, and what is being presented to it,” explains Pestow.

While most collaborative applications involve cooperative coexistence, Level 3 HRC, as much as 10% of applications today incorporate elements of Level 5 HRC, says Pestow.

“I wouldn’t say there is high demand for applications requiring Level 5 HRC, but they do exist, and it is something that we work on in terms of applications and opportunities.”

Human-Robot
Collaboration

Degrees of Separation

Time Together

Sample Application

Level 1

Fenced/Guarded Robot

Human enters the robot’s workspace only under strict safety conditions.

Testing of dangerous substances.

Level 2

Coexistence

No fencing between worker and robot; both have their own workspace.

Pick and place is performed manually to prepare items for a packaging robot.
 

Level 3

Sequential Collaboration

Worker and robot share the workspace but perform tasks at different times.
 

A human lines up product in a tray for a CNC machine tending robot.

Level 4

Cooperation

Robots and humans work at the same time in the same workspace.

A human prepares items for a welding robot and inspects them when finished.

Level 5

Full/Responsive Collaboration

Humans and robots work on the same product at the same time, with the robot responding in real-time to the worker.

Futuristic Level 5 HRC scenarios envisage robots calibrating their behaviours based on humans’ emotional states and stress levels.

A human and robot working together on automotive wiring and inspection tasks.

 

A domestic humanoid robot detects heightened stress levels and offers to provide a destress routine and a drink.


Power and force limiting and speed and separation monitoring features are key enablers of all levels of HRC, enabling humans and robots to work in sequence or together without the need for guarding, following a risk assessment (ISO TS/15066). 

However, as the level of HRC rises and humans and robots get ever closer in the workplace, the potential risk to human workers rises in proportion.

Sensors and safety tools — from area scanners to light curtains and LiDAR — support advanced HRC, but it all begins with the risk assessment, says Pestow.

“I am often asked ‘Is this setup safe?’ The safest answer is ‘What does your risk assessment say?’ Anything is possible, including full Level 5 HRC, but that doesn’t mean that everything will be safe. In addition to a risk assessment, you have to consider how your risk reduction measures are implemented and validated and whether the correct training procedures are in place,” he says. 

Cognitive collaboration

Image courtesy of Universal RobotsWhen cobots first emerged in the mid-2000s, “collaborative” simply meant robots that are power and force limited, explains Chris Savoia, Head of Ecosystem – Americas at Universal Robots, a pioneering company in the collaborative automation space.

Power and force limiting functionality made it possible, for the first time, for robots and humans to share a workspace without guarding. Subsequent innovations in ease-of-use lowered the cognitive burden and in doing so, changed the dynamic from simple physical collaboration between humans and robots to more of a cognitive collaboration,” says Savoia. 

Given the advent of AI and foundation models for robotics, the next step in HRC is to be fully ‘cognitively collaborative’,” explains Savoia. 

“Today, when programming robots, you typically need to put yourself in the mind frame of a robot. You need to think in a linear manner, to understand the right hand rule and what a Cartesian point is. You need to think like a robot in order to program a robot. That isn't the way humans work.”

For Savoia, Level 5 HRC means making robots “truly cognitively collaborative so that they are less of a burden on humans, who won't have to put themselves into the mind space of the robot and can interact with the robot on their own terms.”  

Smarter cobots

AI hardware and software will enable rapid progress in collaborative automation, with technology leaders like NVIDIA developing powerful chips, foundational AI platforms, and simulated training environments that enable rapid deployment of advanced collaborative applications.

“Physics-based simulations allow you to quickly train and deploy AI models. And libraries for robots give them the ability to perform tasks from motion planning to object grasping and vision processing,” says Savoia.

The next evolutionary step in HRC will see cobots with human-like abilities across different application domains, says Alex Lee, North and South America general manager at Doosan Robotics, a leading cobot manufacturer.

Image courtesy of Doosan Robotics“The next phase won’t be so much to do with these pre-defined levels of HRC — we’re already there in many respects. But it will involve AI and downloadable applications. The idea is that you buy one flexible, dexterous robot and then download AI modules that enable it to complete specific tasks safely and collaboratively,” explains Lee. 

Currently, there isn’t huge demand for full Level 5 HRC in the industrial space, says Lee. And with industry trending toward full automation and lights out manufacturing, Level 5 HRC may always be a somewhat niche scenario.

“Manufacturers are not looking for Level 5 HRC as much as they are looking for robots to do the physical tasks that humans are still doing, such as feeding film on packaging lines, palletizing, and operating CNC machines,” explains Lee. 

Futuristic scenarios in which robots take EEG and other readings from humans to guide behavior are “pushing the limits,” says Lee.

“With any new technology that comes onto the market, everybody loves to stretch it. This type of research could help push the next evolution in robotics, but there is no demand for this type of application in industry.”

The humanoid factor

Mention HRC and attention immediately falls on cobots, but there’s a distinct possibility that widespread adoption of “full” HRC will be spurred by humanoid robots more than by collaborative industrial robot arms.

“Twenty years ago, Honda’s Asimov robot showed us that even simple humanoids can function around human beings, while avoiding collisions,” says SICK’s Pestow. “Today we have running, jumping, acrobatic humanoids that are more like ninja warriors. It’s easy to imagine a humanoid robot being used to accompany delivery drivers, for example, and taking care of the physical doorstep deliveries.”

Level 5 HRC will likely take center stage in the service and domestic robot spaces, says Universal Robot’s Savoia.

“The home is where the most robust level of cognitive collaboration will be required because the people involved will have the least training on robots, and you will have to be able to trust the robots around children.”

In the industrial environment, there’s a cost and complexity to humanoid robots that the economics of scale have not yet been able to drive down, adds Savoia.

“For most industrial tasks, collaborative robots will do it all at a lower price point and with more efficiency than a humanoid. But life imitates art and technology imitates science-fiction. We’ve been dreaming about humanoid robots for a long time, so collaborative humanoid technology will eventually come to fruition.” 

Regulatory environment

Safety regulations will have to be updated for domestic robots capable of safe Level 5 HRC to become widespread, but Savoia is optimistic that this won’t prove an impediment to progress.

“Regulations are probably not adequate at this time, but the standards industry in robotics does incredible work and I'm sure they'll step up to the challenge of the future,” he says.

In the industrial environment, current regulations suffice even for Level 5 HRC, but with one caveat, according to SICK’s Pestow.

“Technology always outpaces standards and regulations due to the pace of innovation. The pace that standards are written and updated and revised is always going to be non-linear with innovation in terms of speed and development,” he says.

Collaborative automation has created opportunities for human robot collaboration that were unimaginable just two decades ago, says Doosan’s Lee — and the future is bright.

“I am truly excited about the next evolution of collaborative technology. It will require a complete ecosystem of partners, including hardware, software, and end-of-arm tool companies. The key is that we work together to build practical solutions. This will improve the quality of product and it will improve quality of life for people too. I'm very excited to see where this is headed.”  

Collaborative Robots This content is part of the Collaborative Robots curated collection. To learn more about Collaborative Robots, click here.