Robots have become ubiquitous enough in manufacturing environments that human-robot interaction has become a requirement. It is no longer something that is actively avoided to prevent injury to the human. In fact it is necessary, as programming these types of robots is not always done by a programmer behind a computer. Increasingly, engineers and technicians are standing next to machines, teaching, monitoring or fixing them while other robots hum nearby.
The collaborative nature of these robots has coined the term cobot, and these are quickly becoming the de facto standard for adaptive, responsive and modern manufacturing centers.
Key features of cobots
Cobots are unrestricted by boundaries or fences and, as a result, surpass the borders and workspace restrictions of conventional industrial robotic and automation technology. The primary differences between a cobot and a conventional factory robot are the enhanced safety features that allow operation in close proximity with the human operator and simplified programming techniques that allow simple deployment and redeployment inside a plant.
Various human-robot collaboration (HRC) scenarios are usually defined to ensure a safe working environment. For instance, if a human crosses the border of a defined space around a cobot, it completely stops its operation, allowing the human to safely perform the required tasks.
Furthermore, cobots operate under various speed and separation monitoring and force and power limiting strategies. This means that the closer a human operator approaches a cobot, the slower it moves and at a predetermined threshold, the cobot stops completely. This also detunes the torque used by the robot when a human is in proximity, which limits the impact in the event the human and robot do touch.
Sensing and reception technologies
Cobots must be fully aware of their surrounding and the human presence in it. To achieve this, various exteroceptive sensors are used by cobots to track humans and other obstacles in their environment and form a complete 3D map of the surroundings. Presence sensing can be monitored by one of many technologies, from light curtains to spatial mapping via lidar or 3D vision to RFID tracking. Advanced controllers respond to new risks in the cobot's environment, and reduce speed and torque employed, or stop the cobot altogether, to ensure safe operation.
All cobots employ some combination of exteroceptive technology that simulate key human senses, most commonly sight/vision, hearing and touch. Consequently, the exteroceptive sensors found in cobots generally include vision, touch, hearing, temperature, range finding, acceleration and other similar sensors.
[Discover sensing technologies on GlobalSpec.]
Cobots are equipped with proprioceptive sensors as well, which measure the state of the robot itself. These measurements include wheel and joint positions, motor torques, velocities and other parameters. These sensors are key to data collection and analysis for process optimization and machine maintenance. Data from these sensors is also crucial for controlling the robotic motion and to limit the contact force in case of a collision.
Interoception is another important functionality that is important for safe human-robot collaboration and for self-maintenance of cobots. For instance, in a certain application, a cobot may be required to recognize when its batteries need to be recharged and proactively seek a charger. Similarly, a cobot's ability to detect heat when its internal thermal temperature is too high to work alongside humans is another instance of interoception.
A combination of various sensing technologies, particularly vision and haptic sensors, and advanced programming techniques enable cobots to work alongside humans safely and productively.
Programming cobots
Programming enables a cobot to comprehend the status of its environment and execute dynamic actions. Traditionally, the industrial robots are programmed offline and these programs cannot be modified during task execution, which means there can be iterative debugging. Consequently, a robot operates in a deterministic environment in which the operator does not participate.
Cobots are challenging the traditional means of teaching robots how to operate in an industrial environment. It is still possible to feed explicit instructions to the system via a digital interface, but increasingly this is not the case. Hand-guidance is a feature for many cobot systems, where a human operator physically guides the machine through the desired motion and task, while a computer logs the activity. Some cobot systems also support this via an VR or AR feature and the operator doesn't even need to be physically present at the machine. A third option is that the cobot implements machine learning to learn from both humans and other cobots at the facility.
The net result is a robotic cell that is extremely adaptable and easy to reprogram, which is essential as the software skills gap in the manufacturing sector remains formidable.
Motion control
Many of the mechanical components and motion control systems from a standard multi-axis robotic arm would be implemented in a similar cobot - namely motors, servos, actuators, gear sets, brakes, clutches, sensors and drives.
[Discover machine and motion controllers on GlobalSpec.]
However, due to the need to exert force control, multi-axis torque sensors are often outfitted to the cobot. Whenever the robot detects torque or resistance in an unexpected way, it informs the robotic controller, which ceases cobot operation. These are ideal for tightly controlled environments, where operating parameters are within a known range, measured by a multi-axis torque sensor, and can be programmed into the controller for monitoring.
For this reason, many cobots in adaptive and busy environs may opt for indirect force control, where the cobot has preset stiffness and impedance controls built-in. In effect, the cobot effector acts like a spring, giving way when pushed and pliable to unexpected mechanical input. This has the additional benefit of being a dissipative system for any movement malfunction for the cobot.
Conclusion
In many aspects, the cobot industry is still in its infancy, and business models, value propositions and customers are still developing. Yet cobot technology optimally combines human skills with advanced robotic technology to create a state of art industrial environment that is a key part of Industry 4.0.
Although cobots share common DNA with industrial robots, their proximity to humans requires different strategies and technologies to ensure human works stay safe, and the cobots remain productive.
