Part of my work at Universal Robots involves participating in the continuous work to update and improve the international standards relating to the safety of industrial robots. Most notable are ISO 10218-1 and ISO 10218-2 which define the safety requirements for industrial robots and industrial robot systems respectively.
One of the positive side effects of participating in these meetings is that I get to meet a group of brilliant people from various companies and organizations, all of whom have an interest in robot safety. One of these participants, Dr. Matthias Umbreit, who is now retired from the DGUV in Germany, was kind enough to share with me the statistics of work-related injuries involving robots in Germany. The data are highly credible - the DGUV is a large German organization which is responsible for the insurance of workers in Germany, and it keeps high quality statistics on work related injuries.
The chart here shows the number of ‘notable accidents’ involving industrial robot applications in Germany from 2013 to 2020. By the definitions of the DGUV, a ‘notable accident’ is an accident which results in at least three workdays of absence. To get an estimate of the risk of an accident with a robot application, I also found information about the estimated number of industrial robots in use in Germany in the same years from the International Federation of Robotics (IFR). Based on this, I calculated the average rate of notable accidents per robot for each year, which resulted in the following figure:
One thing which immediately strikes the eye is that the numbers are relatively low. An accident rate per robot of 0.00043 as it was in 2020 means that the average risk of an accident with a robot was 0.043% per year of operation. For comparison the Occupational Safety and Health Administration in the US has estimated that the risk of an accident with a forklift is approximately 8% per year – meaning the risk is 186 times higher than the risk of a robot application accident.
To put the numbers into perspective, we can make a comparison with the number of accidents in general. In 2018, there were a total of 35732 notable accidents with stationary machinery in Germany of which 169 involved industrial robots. This means that robots were involved in less than 0.5% of accidents with stationary machinery.
In the figure I’ve also added a trendline for the number of accidents per robot. This shows an average year-over-year improvement of roughly 6-7%. The clear conclusion is that the safety of industrial robotics has been steadily improving at a relatively high rate. It is a very positive result, and many people share credit. Personally, I hope the trend continues, as even one accident will always be one too many. So, even if there is reason to celebrate the positive results, let us all continue to strive for improving the safety of robot applications even further.
One thing I haven’t touch on so far in this post is robots’ ability to prevent injury. Often robots are deployed to remove human workers from repetitive or dangerous tasks.
Unfortunately it is harder to do statistics on the prevention of injuries compared to the cause of injuries, so quantifying how many injuries have been prevented by the introduction of robots or other types of automation is difficult. However, there are some statistics available which do make it clear that the potential for reduction in the occurrence of repetitive strain injuries alone is significant.
In July 2023 the Center for Disease Control published a report which concluded that approximately 9% of the adult US population had reported injuries from repetitive strain during the last 3 months. Of course the symptoms from repetitive strain injuries vary a lot in severity, but in the same report it was also documented that 22.7% of those who were affected by repetitive strain injuries had symptoms severe enough that they had sought medical advice. Multiplying by the adult US population this leads to more than 5 million people in the US seeking medical advice due to repetitive strain injuries in the last 3 months alone.
It is of course difficult to estimate just how many of the repetitive strain injuries which could have been prevented by robots and automation, but my view is that if we want to reduce the risk of being injured by going to work, we should definitely automate, more not less.
David Brandt, PhD, Technology Officer at Universal Robots.