Five Key Collaborative Robot Safety Concepts

Collaborative robots are power and force limited industrial robot arms that can, when fitted with collaborative grippers and other peripherals, perform collaborative applications from material handling and machine tending to screwdriving and quality inspection.

One of the major robotics success stories of the past decade, collaborative robots can –and often do-- enable the creation of work cells in which humans and robots work together without the need for additional safety guarding.

‘Can’ is the key word here.

While collaborative automation is synonymous with safe operations, merely owning a collaborative robot and a collaborative end-effector is no guarantee that your final application will be safe. Some straightforward but important steps –including a risk assessment— are required to get from safe in theory to safe in practice.

Understanding some key concepts goes a long way towards ensuring safe collaborative automation deployments, so we asked three experts to share their insights on collaborative automation safety topics. Along the way, they provided tips for people new to collaborative automation and busted a pair of popular myths.   

Myth of inherent safety

Contrary to widespread misconception, collaborative robots are not inherently safe. 

“Buy a cobot because you’re thinking ‘I want the flexibility of collaborative applications and I understand that I need to do a risk assessment regardless of whether the application is collaborative’,” says Roberta Nelson Shea, Global Technical Compliance Officer at Universal Robots and a leading expert on industrial robot safety.

The biggest misconception about collaborative automation safety is that “you can just integrate your robot, end-effectors and, in many cases, other machinery such as a CNC machine, without having to think about combined hazards or residual risk,” says Morten Kühnrich, Senior Safety Expert at collaborative automation specialists, OnRobot. “You will still need to do a proper risk assessment and perform risk mitigation on the factory floor before deployment.”

That said, leading collaborative hardware makers invest many design and engineering hours in the development of safety features from collision detection technology and safety rated/monitored stop to power & force limiting systems. In addition, collaborative robots have smooth rounded edges and minimized pinch points along their joints.

Moreover, collaborative robot arms tend to be smaller and more lightweight than traditional industrial robot arms and they typically handle smaller payloads -two factors that reduce energy transfer and the potential risk to human workers in the event of a collision.

Full system safety

Collaborative robot arms are platforms. To perform a specific application, you will need to add an appropriate end-effector, such as a collaborative gripper, screwdriving tool, or sanding disc. Your risk assessment needs to incorporate these elements.

“OnRobot’s grippers have smooth and rounded edges, not just on the gripper housing, but also on the fingertips. Additionally, we design motors that can’t go above certain thresholds in order to minimize the contact pressure if a human is struck by the tool when in motion or gets body parts clamped between the tool and other parts of the workspace,” explains Kühnrich. “In addition, our RG2 and RG6 grippers have safety sensors, so that if you insert your fingers in a spot that you shouldn’t, it will detect it and stop the gripper.”

Apply yourself to the application

Collaborative robots give you the ability to create a safe automated cell where humans and robots share the same workspace, but it is up to you (or your integrator) to make sure the application is safe –and collaborative, says Eric Potter, General Manager, General Industry & Automotive Segment at FANUC America Corporation, a leading player in the collaborative automation space. This means focusing on the requirements of each specific application.

“Your setup may incorporate a collaborative robot, but don’t assume that this means you have a collaborative application. It doesn't matter whether you’re deploying a collaborative robot or not --if your robot is holding a knife and waving it in the air, it’s not a collaborative application. You have to be very careful about not just the robot, but everything else around the robot including what the robot is holding.”

The “great unknown” in any collaborative application is the workpiece, says OnRobot’s Kühnrich.  

“Humans are really sensitive to sharp objects and edges. If you combine with movement at head height it can be problematic. Glass objects, items with sharp edges being removed from CNC machines, metal sheets being moved into hydraulic presses; ensuring safety around these workpieces has to be part of your risk mitigation strategy.”

Don’t fence me in

The second myth busted by our experts is the idea that collaborative automation can always be safely deployed without safety guarding. In many cases it can, but several factors will determine whether your collaborative cell can be deployed cage-free.

These factors include the workpiece, the speed and force with which the robot arm is moving, the payload, and risks involving other machinery, such as CNC machines or presses. Additional risks may be inherent to the application itself, such as tasks that involve handling toxic substances, or tasks that can produce potentially hazardous sparks, such as welding.

The Risk Assessment

Released in 2016 by the International Standards Organization, ISO/TS 15066 is the document that specifies safety requirements for collaborative industrial robot systems and provides guidance on the safe operation of cobots, including an annex for the use of Power and Force Limiting (PFL) which provides details about biomechanical forces and thresholds of contacts with humans.

“It’s important to point out that the clamping pressure threshold levels are at ‘onset of pain’, meaning that if there is a collision with a human, no pain or injury will result,” explains Universal Robots’ Nelson Shea who was on the ISO Committee that developed ISO/TS 15066.

The risk assessment process is “no big deal” for most companies, says Nelson Shea. “It just takes a little bit of time to set up some scenarios and take some measurements and say ‘OK’ (or not). Moreover, there are several specialist companies offering pressure and force transducers for taking measurements.”

Companies that are new to automation should consider hiring a specialist firm to take care of the risk assessment process for the initial deployments, says FANUC’s Potter.

“It’s good practice, especially if you are not familiar with specifications like ISO/TS 15066, to get an expert integrator to step you through that process, certainly for the first 2-3 deployments or until you can confidently handle the process in-house.”

External links:

COVR Project

ISO/TS 15066