Unlike traditional industrial robots which are dangerous, big and inflexible, cobots can operate alongside human workers without needing a fence or a cage. The development of collaboration robots has seen them move to practical reality from the realms of science fiction. However, these intelligent and powerful humanoid machines working collaboratively with humans only exists in the fictional world until recently.
The early models of industrial robotic arms are significantly unsophisticated and feature basic sensors that provide feedback on the joints positions to the control computer. If there is an out-of-range movement being commanded by the controller, these robots are fitted with safety sensors in the form of limit-switches to stop the arm from damaging itself. Early industrial robots are also big and are bolted to the floor. They often wield heavy wielding tools that have great dexterity and speed.
Humans are kept at a safe distance from early industrial robots while its arm is working. Since the 1980s this has been the reality, especially in large-scale car assembly factories. However, their inflexibility and high cost have meant that only a handful of companies can afford them. So, how is technology turning this ‘monsters’ into benign and flexible cobots?
Removing the fences or cages
Roboticists are using simple PIR detectors to sense someone approaching too closely and shut down the machine. This is known as a Safety-Rated Monitored Stop and is quite easy to implement. However, it is equally highly disruptive since the generous safety-margin needed can lead to inevitable shutdowns.
Employing infrared and ultrasonic range sensors around the cobot provides a more measured solution thus it is the better method. Instead, the switch-off zone is replaced by a slow-down zone. What this does is to ensure that the machine reduces its speed if a person is infringing on the safety margin. If the person ensues with their approach, the machine brings itself to a standstill. This method is known as Speed and Separation Monitoring.
Removing the ‘Do Not Touch’ restriction
The term collaboration doesn’t necessarily mean that human and robot need to get close enough to interact. In a collaborative environment, physical contact between human and machine may be unavoidable but indeed desirable which will allow the machine to learn from experience.
For a cobot to achieve its goal, it needs to have eyes, this is where vision systems come into play. There are collaborative robots that are equipped with cameras and can be trained to recognize objects. Recently, tracking systems and object recognition have seen massive improvements in performance this playing an essential role in the advancement of machine vision systems.
Learning or programming?
If you want to train your robotic machine to perform a particular sequence of movement, there are three main options you can use; programming, pendants, and hand-guiding. The Pendant option is only applicable to traditional industrial robots that operate behind cages. This is where a human operator uses a pendant control box located outside the enclosure or fence to manually drive the machine through its normal routine and save the map on memory. This way the machine is taught want to do by its human operator.
A plethora of robots are still programmed with code which is generated off-line on a computer. However, the drawback with programming is that it takes a significant amount of time because your staff needs to undergo training to be equipped to modify or create code each time the cobot needs to be re-deployed. This rather defeats the purpose of a small company owning one.
Hand-guiding just like Pendant uses the principle of teaching except that the former is hands-on. With this option, your cobot is taught to mimic a particular action through moving its arm in the exact sequence required.
Artificial intelligence is set to be the dominant feature of robotic control in the near future. This means that cobots will learn for themselves how to achieve a specific goal.