We have kept our robots locked away behind protective guarding for decades, working in areas separated from humans. Safety gates and presence sensing devices monitor the robotic work cell and are integrated into the Safe Off functions of the robots, ready to place the robot in a safety state when a person is detected in the robotic work cells.
We did create shared spaces where a human and a robot take turns entering the same area. A robot may place a part into a fixture and leave the area; then a person enters and places another part over the first. They leave and the robot returns to weld the 2 parts together, pick up the part and move it to the next fixture. This approach has sophisticated functional safety detection mechanisms for both the humans and the robots in the shared space.
These traditional hard guarding approaches will always have their place. However, a new wave of robots require more freedom and more human interaction. This requires an order of magnitude more sophistication from the functional safety system.
Automated forklifts and Autonomous Mobile Robots (AMRs) are sharing warehouse and factory floors with workers. Humanoid robots are going to be in public spaces. Robots are being developed to move your packages from the delivery vans to your front door. These systems require mobility and interaction in human occupied spaces.
More sophisticated safety systems enable robots to be aware of their surroundings, detect humans in their vicinity, and respond appropriately. The standards to support this shift are emerging from the best practices that are currently being established. We need to ensure that when robots are expected to operate outside of protective work cells, they still operate with the appropriate amount of safety.
This represents a fundamental shift in robot safety systems. These robots operate in dynamic environments where conditions change continuously. The challenges with this change are significant.
New Safety Requirements: Awareness and Response
Situational Awareness
For robots operating in areas with humans, situational awareness becomes the foundation of safety. These systems must continuously perceive their environment in real-time, understanding not just where obstacles are, but which obstacles are human and their possible path of movement.
This goes beyond simple presence detection, it requires understanding human behavior, recognizing different types of human activity, and assessing the level of risk associated with different scenarios.
Obstacle identification extends beyond humans to include other robots, vehicles, equipment, and environmental constraints. The robot must understand the context of its operating environment.
Appropriate Responses
Awareness alone is not enough. Robots must also respond appropriately to the situations they perceive. Speed and separation monitoring ensures that robots maintain safe distances from humans, adjusting their speed based on proximity and the ability to stop safely if needed.
Safe stop capabilities are fundamental. When elevated risk is detected, the robot must be able to stop quickly and safely, but not every detection requires an emergency stop. The system must be able to assess the level of risk and respond proportionally: slowing down, changing path, or stopping completely based on the situation.
Collaborative operation modes enable robots to work alongside humans when it's safe to do so. These modes might involve reduced speeds, limited forces, or restricted motion ranges that allow safe interaction while still enabling productive work.
The Functional Safety & AI Standard
New standards are emerging to address the unique challenges of modern robot applications. The ISO/IEC TS 22440 standards address the integration of AI with Functional Safety systems, recognizing that many of these new robots rely on artificial intelligence and machine learning for their perception and decision-making capabilities. This standard is critical because it addresses how to ensure functional safety when the system's behavior is determined by AI algorithms that may not be fully predictable.
For safety engineers and robotics professionals, understanding this standard is essential for safe deployment of modern robot systems. The robots are being freed from their cages. Our safety approaches must evolve to keep pace. Staying current with the new standards is part of the job for anyone designing or approving systems that free robots from the cage.
The future of robotics safety lies not in preventing interaction, but in enabling safe, productive coexistence between humans and robots in shared spaces.