When Industrial Robots are Dressed as Cobots

Safety is a pillar of smart manufacturing, especially when it comes to human-machine interaction: how long humans and machine should work together and how they should do it while keeping an excellent production level are certainly the two most important factors. In addition to industrial robots and cobots, Mitsubishi Electric has designed intermediate solutions: an industrial robot capable of becoming collaborative as needed, without affecting the industrial process or jeopardizing operator safety.

by Rossana Pasian

Safety is one of the three pillars that make up Mitsubishi Electric’s robotics philosophy, perhaps the most popular one among competitors, speaking of the growing need to make humans and machines interact according to the percepts of the most advanced smart manufacturing. However, Mitsubishi Electric’s approach has been slightly different from that of the market, having implemented a gradual and measured policy to the real application need. Having clearly understood the objective of making humans and robots collaborate and work side by side, they wondered how to achieve this without necessarily involving a cobot, but also using other devices and tools to make the production scenario collaborative.
“While everyone was talking about collaborative robots,” says Marco Filippis, Product Manager Robot South EME at Mitsubishi Electric, “Mitsubishi Electric, became a representative of the collaborative application concept by implementing a technological approach with ‘hybrid’ solutions, which is scalable according to the real application, up to the actual cobots. Different solutions have been developed based on the interaction time between machines and humans”.

Industrial robot vs. collaborative robot
If you want large productivity at the end of the line, you need to use an industrial robot, that is the “classic” robot that has been seen for many years in the factories, which must necessarily be closed inside a protective cage, as it reaches particularly high speeds with the risk of hitting the operator. On the other hand, there are pure collaborative robots, which can coexist with humans without any kind of barrier; however, for safety reasons and for the robot nature, they cannot guarantee equal speed and productivity. Between these two, which can be defined as opposites, there are several intermediate solutions, capable to resolve some of the users’ problems.
“The issue of Industry 4.0”, explains Marco Filippis, “was to let the public perceive two concepts that in my opinion are wrong. The first is that the collaborative robot would have been the future of industry in any field, but this is not possible: indeed, the cobot has a very clear nature and a well-defined position in the market; if you use a collaborative robot for industrial applications, at some point there will be shortcomings regarding speed or mechanical structure, because if you work beyond physical limits there is a risk of breakage”. “The second mistake” continues Marco Filippis “is that there has been binary, if not conflictual, management between the industrial robot and the collaborative robot. But these two solutions are designed to perform different operations, and one cannot replace the other, at least in industry as it is now”.

“Cooperative” applications
Meanwhile, in the middle there are upcoming applications that can be considered “cooperative”. At the dawn of Industry 4.0, Mitsubishi Electric had already developed the MELFA Safeplus advanced safety module. It is a control unit which is connected to the robot controller and Safety elements, from barriers to laser scanners up to eight Safety inputs. Thanks to external sensors, it is possible to create collaborative areas around the robot. This means that they go at different speeds: low during the approach and zero when the operator is close to the robot. The second step of MELFA Safeplus is the MELFA Safe Skin: in addition to the Safety inputs delimiting the warning area, industrial robots are covered with a multisensory “skin”: sensors that touch the robot in order to stop it.
The great advantage of these hybrid solutions is that you are not forced to affect productivity at the end of the line, even if the operator is far away, and at the same time it is not necessary to stay away from the robot for fear of hitting people.
So, it has been developed an industrial solution capable to become collaborative when needed: “We dress an industrial robot as a cobot” comments Marco Filippis. The advantage of this solution is that you have certified functions. But, above all, by creating collaborative areas, you no longer have to worry if what is around the robot is collaborative or not: it is an industrial machine that is made collaborative within a defined value.

Collaborative robotics approaches industrial robotics
Mitsubishi Electric has also developed the MELFA Assista robot. Released in 2020, it provides a payload of 5 kg on the wrist and a reach of 932 mm, thus positioning itself, in the field of collaborative robotics, as an intermediate product. After its marketing, work will begin on the robot scalability, in order to reach the 10 kg payload and have a complete range that meets a wide range of applications.
MELFA Assista has fairly defined characteristics: it does not have batteries for the motors (it uses Mitsubishi Electric’s MRJ5), and boasts a very high repeatability, of about 0.2 mm in position, because, even if the internal structure is collaborative, they have chosen to keep some details that are characteristic of industrial robots. The mechanical structure on the last two axes is very rigid, which allows for a repeatability comparable to their industrial “brothers” both in direct teaching, that is when the robot is dragged to specific positions, and when it is used via software.
The cobot software is easy-to-use, depending on the fact that collaborative robots often end up in the hands of end users, who do not have a robotics background. MELFA Assista, as mentioned, can be programmed, in addition to direct teaching, using a programming keyboard with two alternative software, both from Mitsubishi Electric, RT ToolBox3, which covers all robots, or the special software RT VisualBox. With the second, it is possible to create a flowchart via the app; this allows the user to manage the way in which the axes are connected, manage conditioned branches (if it is a single branch, two or more branches) to define what the robot must do at that moment, the opening and closing of the grippers and much more. The advantage is that it is all very graphic, and consequently direct. If you need to do something more complex, switch to RT ToolBox3, which gives you more flexibility. In 3D simulation it is possible to import CADs and, with a good approximation, define the work cycle of the robot.

Applications for cobots and intermediate solutions
“I believe there are applications designed to use a cobot,” comments Marco Filippis. “These applications provide a close link between man and robot such as coordinated operations in which the cobot performs operations that make it a co-worker.” There are repetitive, wearisome and demanding activities that do not necessarily have to be performed at high speeds, because the operator must be given time to do their part of assembly or handling: for this type of applications the cobot is ideal. Hybrid solutions, on the other hand, include all those applications in which it is necessary to reach a productivity that is very close to the industrial one, but at the same time the operator can occasionally enter the work area. Consequently, the choice between the use of a pure cobot or a hybrid solution is dictated by the time of interaction with the human.
“Maybe in the future there will be new applications dedicated to the cobot,” says Marco Filippis, “already designed to use it. Mitsubishi Electric believes in the Japanese philosophy of ‘small steps’: when one of our products comes out, we know that it is tested and guarantees excellent quality and interesting performance for the market. Mitsubishi Electric is known for quality: we cannot allow customers to say that one of our products does not meet the standards. We base our corporate strategies on this”. For MELFA Assista, it went exactly like this: it was developed when this type of hybrid application was not yet well known, it was refined and then presented to the market.

Safety and artificial intelligence
Artificial intelligence also plays a key role in safety and in the coexistence between robot and human, facilitating the operator in their work and making the robot smart. In fact, the robot itself is created as an enslavement arm, if it is not integrated with external sensors telling it what to do. Mitsubishi Electric, by using AI, intends to improve the robot performance in terms of handling both in relation to humans and in somewhat particular applications.
With MELFA SmartPlus, Mitsubishi Electric offers solutions, hardware plus software, capable of improving the functions that the robot already has, such as the use of the force sensor. This is useful in all those applications where the robot must adjust to a non-linear path or to a very precise assembly. Mitsubishi Electric has integrated an artificial intelligence algorithm with the force sensor, which allows you to change the robot speed control: in this way you use deep learning, because basically what you have is a detection and modification of the speed control in order to achieve a result that improves speed and efficiency. The robot then learns to adjust its speed and strength according to the situation.
Furthermore, Mitsubishi Electric has invested – in collaboration with Real Time Robotics – in the project to create the Robot Motion Planning application, which helps to avoid possible collisions between robots and humans or between robots and robots, and makes it possible to stop robots without touching them. The robot understands the dynamic variations around it and then dynamically corrects its trajectory in real time. All this is possible thanks to a camera system and dedicated hardware, together with the artificial intelligence algorithm, which allow you to work in a completely deconstructed system. “It is the robot that adjusts to the presence of humans and not the other way around” concluded Marco Filippis.