A cobot, or collaborative robot, is a type of robot optimized for collaborative applications. Working collaboratively is the pinnacle of human robot human as described in my previous blog in this series, see here. When people think of collaborative robots, they typically think of something like the UR5, Kuka LRB series, Fanuc CR-35iA, ABB YuMi etc. Cobots as a form of industrial robot are distinct from service robots in that they are designed for industrial environments, to build things, as opposed to assist people in their everyday lives, but in both cases, they are designed to interact with humans.

Typically, when people think of Cobots they think of robots implementing Power and Force Limiting (PFL) which I hope to describe in my next blog in this series. However today I will focus on Speed and Separation Monitoring (SSM) which is defined in sub-clause 5.10.4 of the industrial robot safety standard ISO 10218-1:2011.

Figure 1 –  Description of SSM from ISO 10218-1:2011

With SSM, the area around the robot is monitored and if someone approaches the robot it will initially slow down and eventually stop. The goal being to ensure that the robot is stopped before contact can be made with the human or part of the human. The advantage of SSM over PFL is that the robots should be able to run faster, which increases productivity. PFL speeds must be slow to keep the forces low in case of collisions. With SSM, the robot does not need to stop immediately someone enters an area but instead can monitor the person, their direction of travel and either change its poise to keep further from them or slow gradually as they come closer. Eventually the robot/cobot might have to run as slow as if it was using PFL so that the two modes can in fact be complementary. Slowing is generally (I was going to says always but I’m sure someone would come up with a corner case to prove me wrong) more efficient than stopping. Then as the person moves away the robot can start to move faster again. Always remember that the goal on the factory floor is achieving the highest possible productivity allowing for the need to maintain safety.

The picture below shows an example where at a distance up to the green line the robot can operate a full speed before needing to come to a complete stop before the person reaches the red line. In between the robot begins to slow. A warning might be issued at the orange line, but the Output Signal Switching Devices (OSSD) go active at the red line and the robot should stop or enter PFL mode.

Figure 2 – Picture above shows zones where the robot can operate at full speed, reducing speeds and when it should be stopped

The math behind the distance between the lines is shown below. It considers things like how long it takes the robot to stop, how fast the robot is moving, the mass of the moving parts of the robot, the length of the robot arm etc. Typically, it is assumed that being on a factory floor no running is allowed and a maximum walking speed of 1.6 meters/second is assumed.

Figure 3 –  Protective separation distance calculations based on ISO TS 15066

The graphic below from ISO TS 15066 shows the practical impact of the above. While ISO TS 15066 is currently a companion technical specification to ISO 10218, revision 3 of ISO 10218 is due out in late 2022 and the contents of ISO TS 15066 will then be within the ISO 10218 standard.

Figure 4 – ISO TS 15066:2016 Figure 3

Further guidance is also given in ISO 13855.

ISO 10218-2 requires that the sensors (ESPE and AOPD) used to implement SSM shall comply with the relevant parts of the IEC 61496 series. In a future blog in this series, I will discuss the IEC 61496 series but for now you just need to know that IEC 61496-2 covers light curtains, IEC 61496-3 covers laser scanners and 3D TOF.

Figure 5 –  Extract from ISO 10218-2:2011 showing call to IEC 61496 series of standards

Obviously using laser scanners, light curtains, and 3D ToF sensors are not suitable for all applications. For instance, if something can be ejected from the robot cell then some sort of physical guard or fence might be better.

If you are using electro-sensitive protective equipment (ESPE) then 3D ToF can have advantages compared to laser scanners. For instance, laser scanners only operate on a single plane. Typically, with laser scanners, the monitored plane is close to ground level so that the laser scanner cannot detect an outreached hand but can detect a body on the floor. A 3D ToF camera, however, monitors in three dimensions and can mean you don’t have to allow an extra 0.85m for the outstretched hand. Typically, however the range and FOV (field of view) for a laser scanner is bigger (flash vs point illumination). Having a 270-degree FOV for a camera is difficult but common for a laser scanner. But increased FOV comes with reliability issues due to the moving parts within a laser scanner. A camera-based system has no moving parts. For mobots, the 3D monitoring can be used to detect overhanging objects which a laser scanner could miss and potentially the data from the 3D monitoring could also be used for object classification so that a robot could behave differently in the presence of other robots as opposed to people. Object classification being made possible by the rich data generated by the camera. Both the camera and laser scanner systems while there for safety can also be used to implement non-safety functions such as navigation.

Regardless of the technology involved (scanner or camera) a typical requirement is PL d according to ISO 13849, SIL 2 according to IEC 62061/IEC 61508 and type 3 according to IEC 61496. Most solutions are specified to all of these.

For more information on ADI sensors suitable for 3D ToF cameras see here. A future blog in this series will cover 3D ToF cameras in the context of their use in Cobots and Mobots (mobile robots). Most laser scanners also contain lots of ADI components, so I guess ADI wins either way, but it good to know the issues for our customers and our customers customers.

There is a good demo of SSM available here.

For all previous blogs in this Safety Matters series, see here.

Note – ISO 10218 series is also available as ANSI R15.06. A new version of ISO 10218 will hopefully be published later in 2022 and will include ISO TS 15066.

Note – I have deliberately been sloppy in the use of the terms robot and cobot because there is a really no such thing as a cobot. A cobot is simply a robot with features which make it suitable for use in collaborative applications.

Interestingly a lot of the same technology described in this blog is also relevant for automotive. However, for automotive the ranges can be up to 200m as opposed to 5m, the cars are moving faster than in the robots and in a more uncontrolled environment. On the factory floor it is for instance assumed that children are excluded and that operators are fully trained.

Anonymous