Manufacturing environments are busy, and avoiding collisions between robots and operators is a high priority. As more manufacturers add robots, there’s increasing interest in ensuring they work safely with each other and with people.
The robotics industry can take pride in its impressive safety record with more than 1.5 million industrial robots operating worldwide, according to Carole Frank, safety director for the Robotic Industries Association (RIA). As robotic applications increase, it’s vital to continue to be vigilant about robotic safety. In fact, risk assessment is now required by new safety regulations: ISO 10218-1 and -2 delineate safety requirements for robots, replacing ANSI/RIA R15.06.
Many robots are certified by a third-party source or approved by their manufacturers. That’s good, but it’s also important to be sure the robot is safe in its surrounding environment. So take a holistic approach and evaluate each industrial application rather than each device separately.
Basically, you want to identify risk sources, estimate the risk, evaluate it, and determine if the risk is acceptable or needs to be mitigated. A risk tree can help you rate each of these parameters: severity of injury, frequency of exposure to hazard, and the possibility of avoiding a hazard. If you’re unfamiliar with the detailed, iterative process of a risk assessment, a systems integrator can provide this service.
Often, when people talk about robot collaboration, they focus on the robot and ignore how the robot functions in its environment and with people. There are four types of collaborative operation:
- Safety-rated monitored stop allows the operator to interact with the robot when it is stopped, and operation automatically resumes when the person leaves the collaborative workspace.
- Hand-guiding operation allows the operator to have direct contact with the robot, using hand controls.
- Speed and separation monitoring delimits different safety zones, so the robot’s speed changes depending on the zone. A protective stop is issued when an operator is in potential contact.
- Power and force limiting occurs when incidental contact between a robot and a person will not result in harm.
Let’s focus on speed and separation monitoring, since it enables the robot to sense the presence of an operator and adjust its speed. Often, this applies to uncaged robots that use safety sensors. At Concept Systems, we create work cells using lasers or a vision system that predicts a collision before it happens.
This approach allows the robot to continue to function while operators are near because the robot works at specific speeds determined by preset safety zones, ensuring human and machine safety. When an operator comes too close, the robot will stop. We often refer to this as dynamic collision avoidance because it underscores the ability of the robot to react to changes in its environment and increases flexibility and utility.
It is often critical that the robot not collide with a large part or piece of machinery. We’re working on a project that does that by using speed and separation monitoring. An operator can observe the robot’s movement at an HMI. When the robot is operating safely, it appears in green on the screen. When it enters a warning zone, the color changes to yellow and the robot slows down. If the robot enters a collision zone, it turns red and stops operating.
Many robots are certified by a third-party source or approved by their manufacturers. That’s good, but it’s also important to be sure the robot is safe in its surrounding environment.
Lidar, the technology used in many autonomous vehicles, has become more common in industrial environments with systems provided by Sick and Velodyne. Using laser light to detect distances, lidar generates exceptional data for navigation algorithms and typically has a longer operating range than other vision technologies, which is critical for collision avoidance when there are changes in the environment, including interactions with people.
Lidar sensors can be selected to provide the degree of data accuracy needed. 2D sensors could be fine for an application that requires vision of a plane, while 3D sensors provide a more complete visual image. When selecting a lidar system, consider these factors: how far it needs to see, the degree of light sensitivity, what angular resolution is needed, the field of view, refresh rate and accuracy.
When existing robotic systems are deployed or modified, regulations now require that a risk assessment be performed. From the perspective of the company, it’s good business to gain these insights into the safety of robot operations to avoid worker injuries or damage to the robot or other machinery.
With the increasing capabilities and complexities of technology, putting a priority on industrial safety is more important than ever. As systems integrators, we discover unsafe situations and help companies mitigate their risks. Risk assessment combined with corrective action has proven to produce a significant return on investment by avoiding unnecessary collisions.