Robots in the workplace
As new technologies bring robots side by side with human workers, what are the safety implications?
In popular fiction, robots have long labored alongside humans, sometimes harmoniously and sometimes disastrously (depending on the requirements of the plot).
In reality, industrial robots have been in American workplaces for decades. The Robotic Industries Association estimates that in 2017 more than 250,000 such robots had been installed in the United States. These machines bear little resemblance to their science-fiction counterparts, usually taking the form of powerful, heavyweight automated arms that perform tasks such as welding, painting or assembly within the confines of a cage or other enclosure.
Recently, however, technological advances have begun to allow for greater diversity of robotic systems in the workplace.
“Now, in addition to traditional industrial robots, we have professional service robots, collaborative robots who work side by side with workers, and mobile autonomous robots in a wide range of industries and enterprises,” said Vladimir Murashov, senior scientist in the Office of the Director at NIOSH and a member of NIOSH’s newly formed Center for Occupational Robotics Research.
As these next-generation robots open new possibilities, their increasing interactivity and mobility may complicate the task of ensuring the safety of their human co-workers.
New roles for robots
Industrial robots typically have been used for tasks considered undesirable for human workers – what RIA Vice President Bob Doyle described as “the three Ds: dull, dirty and dangerous jobs.”
In this capacity, automated systems offer considerable safety benefits to human workers. “Robots can help prevent injuries or adverse health effects resulting from working in hazardous conditions,” Murashov said. “Some examples are musculoskeletal disorders due to repetitive or awkward motions, or traumatic injuries (for example, in poultry processing, where cuts are common). They can also prevent multiple hazards in emergency response situations such as chemical spills.”
Besides keeping workers out of harm’s way, robots also can minimize risks stemming from human error. “If a job is repetitive and boring, human workers tend to make a mistake, whereas robots can do these things the same way time after time,” said Frank Hearl, NIOSH’s chief of staff and a member of CORR.
Next-generation robots still perform these types of tasks, but their ability to work in close proximity with humans creates the potential for a wider range of applications. Interactions between humans and collaborative robots can involve handing off parts and materials, or the worker “teaching” the robot by guiding it through a desired motion, which the robot then repeats.
Employers are just beginning to explore the possibilities for strategic human-robot partnerships. “Collaborative robot systems allow for partially automated tasks where the robot and human can both use their own strengths to the best effect,” said Carole Franklin, RIA’s director of standards development. “You have the strength, precision, endurance and repeatability of the robot, and you also benefit from the flexibility and sensitivity of human touch, as well as human problem-solving and creativity.”
To illustrate this concept, Murashov used the example of a car manufacturing facility where humans and robots work together to insulate and water-seal vehicle doors. “In that case, the robot spreads out and glues down material while the worker holds it in place with more agile human fingers,” he said.
In some instances, collaborative robots are filling jobs left vacant by human workers. This is the case for Dynamic Group, a Ramsey, MN-based plastic injection molds and molding company that uses five robots – two traditional industrial robots and three considered collaborative.
“We have some jobs that just aren’t very fulfilling or attractive,” CEO and Co-Owner Joe McGillivray said. “They’re repetitive, boring – things no one really wants to do, no matter how much you’d be willing to pay them.”
Dynamic Group’s robots perform tasks such as tending machines and handling parts, picking them up and setting them in front of human workers. McGillivray is pleased with the ease of programming and deploying the robots, and he expects the company to increase its use of collaborative robots.
“I have work getting done on all shifts that wouldn’t be done otherwise,” he said. “The robots don’t quit, find a better job or get sick; they’re able to keep going and going. People on the shop floor love them. They don’t want to be doing the jobs the robots are doing.”
What makes a robot ‘collaborative’?
A robot is considered “collaborative” when its design incorporates built-in safeguards that allow it to interact with humans. In its most recent technical specification (TS 15066), the International Organization for Standardization requires collaborative robots to use one of four safety measures:
Safety-rated monitored stop. “This is similar to the approach used with traditional industrial robots,” said Vladimir Murashov, senior scientist in the Office of the Director at NIOSH and a member of the NIOSH Center for Occupational Robotics Research. “The work stops when a worker would like to enter the workspace of the robot.”
Hand guiding. The robot moves only under an operator’s control.
Speed and separation monitoring. The robot automatically decreases its speed as a worker approaches, to the point where it stops when it’s about to be touched.
Power and force limiting. “This is the approach that has really taken off in the marketplace, so it defines what everyone considers a ‘collaborative’ robot,” said Carole Franklin, director of standards development at the Robotic Industries Association. “Typically, these robots are limited in the payload they can carry and in the amount of force they could exert if they were to strike a person by accident. Also, the power/force-limited robots tend to be designed with rounded edges and softer surfaces that reduce the risk of injury if contact were to occur.”
Evolving safety standards
Since their introduction, industrial robots regularly have been designed to operate at a distance from workers – and for good reason. These machines pose a variety of hazards and often lack the sensory capabilities necessary to detect nearby humans.
In 1984, responding to the death of a die cast operator who was pinned between a hydraulic robot and a steel pole, NIOSH published an early set of safety strategies, most of which depended on keeping workers at a distance during operation.
“Traditionally, the safety measures have involved physical barriers, sensors and other techniques to either keep people away from the robot while it’s working or shut it down when people approach too closely,” Franklin said.
What happens, then, when cages and fences no longer are necessary because the robots are designed to be interactive, or even mobile? RIA has been working with the American National Standards Institute and the International Organization for Standardization to keep up with these changing realities. Current safety standards (ANSI/RIA R15.06 and ISO 10218) have begun to account for the introduction of collaborative robots, and a recent technical specification (ISO/TS 15066) addresses the safety considerations involved in establishing a collaborative robot system.
Government organizations are doing their part. NIOSH responded to the growth in workplace robotics by forming the Center for Occupational Robotics Research in September 2017. The next month, RIA signed a partnership agreement with both NIOSH and OSHA to share expertise, raise awareness and direct research to address these new safety challenges.
Building safety into the system
Because collaborative robots are intended to share workspaces with humans, safety measures are built into their design. However, despite these inherent protections, collaborative robots still pose significant risks.
“One of the most common myths that our standards community is trying to dispel is that a collaborative robot is safe out of the box,” Franklin said.
The manufacturer has control over the robot’s design but not how it’s configured (a robot arm from one manufacturer might be combined with an end effector from another, for example), programmed or used.
“It’s very important to distinguish between the collaborative robot and the collaborative robot system,” Franklin said. “The power/force-limited robot arm could be integrated into a system that’s not really appropriate for collaborative use. No matter how soft or rounded the robot arm is, if its end effector or its work piece is sharp, a person could still be injured.”
The environment each robot is placed in and the specific tasks it performs present additional risks.
“One thing that’s easy to overlook is interaction with other pieces of equipment,” McGillivray said. “Our plastic injection molding presses exert several tons of clamping force each time we make a part. If the robot is between those clamps when they close, the machine is going to pick the robot up and throw it across the room, and someone is going to get hurt. Those risks are easy to anticipate and mitigate, though, once you get used to that mindset.”
Because every collaborative robot system is unique, risk assessments are crucial to safe and successful implementation – and a core requirement of the current safety standards. Franklin suggests that risk assessments be conducted before, during and after installation, and considers these initial steps the responsibility of the integrator who puts the system in place. Although large companies may have their own in-house integration teams, smaller firms typically work closely with an outside integrator, according to Franklin. “The integrator is far more than just a pass-through sales company,” she said. “They are very much involved in the design of the robot system that ultimately gets installed and used.”
Although the integrator is responsible for initial risk assessments, RIA recommends that in-house personnel be heavily involved. “That could include the plant safety professionals and even the operator,” Franklin said. “Operators have insights into how they are going to use the system, and they may be able to envision risks that someone else cannot.” Organizations also must keep in mind that if they modify a robot system after it’s in place, it is their responsibility to conduct new risk assessments.
Risk assessments for collaborative robots are similar to those for traditional industrial robots, said Franklin, who suggests a task-based approach. “Installation, maintenance and clearing a jam, for example, are all separate tasks with different risk profiles. You identify a task-hazard pair, assess the level of risk that task-hazard pair presents and then determine what mitigation techniques you should apply to reduce that risk to an acceptable level.”
Robot mobility is a game-changer, and Franklin admits that standards-makers continue to wrestle with its implications for risk assessment.
“What we’ve been dealing with so far is a robot that is fixed in place, so a person has to approach the robot in order to be exposed to its hazards,” she said. “A mobile robot could be moving more or less freely (depending on its capabilities) throughout a workspace. You as a human could be standing there doing some completely unrelated task, and the robot could approach you. We are still considering whether the task-based risk assessment makes sense in that scenario.”
Types of robots
Here are some of the robot types likely to be found in professional settings:
Industrial robot: An automatically controlled, reprogrammable, multipurpose manipulator, programmable in three or more axes, which can be either fixed in place or mobile for use in industrial automation applications such as manufacturing, inspection, packaging and assembly.
Professional service robot: A robot that performs commercial tasks outside of industrial automation applications. For example, a robot used in cleaning, delivery, firefighting or surgery.
Mobile robot: A robot that can travel under its own control.
Collaborative robot: A robot designed for direct interaction with humans.
Source: International Organization for Standardization
Gaining human co-workers’ trust
Introducing new technology in the workplace often is met with a degree of anxiety or skepticism. Employers thinking of adopting collaborative robotic systems should bear in mind that workers may have concerns about their safety or job security, and dispelling those fears can take time.
“It’s the same kind of trust one builds in any co-worker,” Hearl said. “At first, you’re watching the new person perform, and then as you see that they’re doing their job correctly and safely over and over again, you build trust.”
One way to accelerate this process is to show workers how a robot can benefit their safety or job quality. When McGillivray introduced his organization’s first collaborative robot, he was conscious of the need for employee buy-in and careful to involve his workers in the implementation.
“I was very clear about what the robot was here for and what we were trying to do with it,” he said. “People got used to seeing it on the floor, and I let them provide direction for where it could and should be used. We never said, ‘We’re trying to take your jobs.’ We said, ‘We’re trying to give you better jobs,’ and that’s exactly what we’ve been able to do.”