by Vincent Duchaine, Professor at Montreal ETS , Samuel Bouchard, CEO at Robotiq
The first application area for industrial robots was in automotive production plants. Since then, improvements of robots have been intimately tied to the challenges of the automotive industry.
While just a couple of decades ago car assembly lines were an infinite repetition of the same model, today’s lines are lot more versatile. One example of automotive line diversity is at Lansing Grand River Assembly (LGA) located in Lansing, MI. Opened in 2001, it is one of the most modern GM plants that produce vehicles whereby an SUV can follow a wagon and a coupe on the same line. What is the next step? The car makers’ dream is to make and ship a car to a customer’s order in two weeks, just like Dell does with laptops.
This mass customization approach would solve important inventory issues and provide the highest satisfaction to customers to ensure they get their “dream” car. This naturally imposes engineering challenges on automated manufacturing systems. Robot flexibility will need to reach yet another level to account for constantly changing assembly tasks.
Bringing the next generation of robot manipulators into the same environment as humans seems a natural evolution in the path towards more advanced robotics. This co-existence will offer a tremendous potential to improve many assembly line tasks. A more robot / human approach will combine the human’s reasoning ability and adaptability with the inexhaustible strength and precision of robots. This paradigm shift that will bring more flexibility to the production line is ready to happen.
The capability of understanding human intentions and reacting to them is something missing in the current generation of industrial robots but at another level, future cooperative robots will need to be designed in such a way that human safety is guaranteed.
Industrial robots being almost sensorless and built for strength could easily injure a human and thus are forced to operate in confined workspaces. This safety issue partially explains why actual interaction between human and machine on automotive assembly lines relies principally on passive devices. By nature, these mechanisms are not able to supply energy and therefore the risk of injury becomes minimal.
The principal advantage of passive devices is that they can greatly reduce the effort level of a worker needed to move a given object by counterbalancing a static load. This is typically achieved by the use of a pneumatic actuator. Knight Global is a leader in this field and proposes a wide range of solutions like passive arms, suspended cables or torque tubes. Despite being safe, reliable and a good assistive help for workers, passive devices have some limitations. The fact that they cancel only static forces makes them more suitable for lifting small and medium payloads. In this specific context, the force associated to gravity is the only one that is significant but for payloads over a certain weight, friction and inertia forces become more dominant creating forces that could go over ergonomic standards.
Intelligent Assist Devices
In response to the above mentioned limitation, Intelligent assist devices (IAD) have started to make their way to automotive plants. These are actuated versions of passive devices and can be defined as specialized cooperative robots. By using force sensing and advanced control algorithms, these devices are capable of accounting for dynamic as well as static forces. However, despite being able to cancel all the inertia, they typically keep a certain virtual dynamic at the point of interaction so their control remains intuitive enough for humans.
In addition to carrying a load in cooperation with a human operator, IAD’s can also do a part of the task autonomously like picking a load in a similar way as a conventional robot would do. One of the first introductions of IAD’s in automotive assembly plants was a device commercialized by Cobotics (today sold by Stanley Assembly Technologies) that supported workers in the insertion of instrument panels; a task known for requiring both significant force and dexterity simultaneously. Many IAD’s are available in the market today. Gorbel proposes a wide range of products in this area from intelligent lifting devices (suspended cable) to devices similar to robot arms.
Although they do not fit in the generally accepted definition of an industrial robot, the fact that they are powered by servo motors and that they supply energy, makes them a potential safety threat to humans. As mentioned above, this issue is the main reason current robots do not share the same workspace as humans. Indeed, many studies have investigated this particular aspect in a way to demonstrate that a robot could easily injure a human being and also to understand and provide a metric relative to the level of this threat.
Having a human and a robot carrying a load together is already possible with an IAD. The potential benefit of having a human and a robot in the same workspace on an assembly line should go further than this. Other examples of applications could be one feeding the other with a variety of different parts, holding parts so the other can work on it, teaching the robot directly by holding it, etc. Although these potential applications fall within the spectrum of current robot platform abilities, this interaction is impossible to obtain with current industrial manipulators due to safety issues.
Recognizing that industrial manipulators need to be redesigned if we want to use them in close proximity with humans, Kuka recently put their lightweight arms on the market. Based on the LWR3 build by the Deutsches Zentrum für Luft- und Raumfahrt (DLR), this new robot has been proven by researchers to be inherently safe for humans. It is equipped with joint torque sensors and compliant actuators that allow it to sense collision and absorb part of it. Even more important, the power of this robot combined with its compliant property and sensor capability ensures that it can’t produce sufficient force to injure a human in the case of clamping.
More alternatives have made their way to market recently or should be available soon. A few examples are:
- The human-friendly manipulator recently commercialized by Universal Robots
- FRIDA the new concept Friendly Robot for Industrial Dual-Arm of ABB
- A low power safe robot that is currently used in Toyota assembly plants that has been built for human-robot interaction
- The Workerbot from German pi_4 Robotics
- The product from Boston start-up Heartland Robotics founded by iRobot and MIT veteran Rodney Brooks that promess to be “teachable, safe and affordable”
Flexible Robots, Flexible Peripherals
Increasing flexibility in automated manufacturing will definitely require more than advanced robot manipulators. These new platforms will need improved programming, sensing and tooling to cope with the high variety of different parts and operations.
The fact that an operator will be able to interact physically will enable new way to teach control points without the teach pendant. A recent demonstration of programming a massive Kuka robot using a joystick mounted on it at Automate 2011 gave a preview of such a programming approach.
More intuitive vision hardware and software will be necessary to adapt to the task complexity. In that sense, 3D vision packages based on 2D cameras that are easy to teach are being brought to market. Examples are Comau’s Recognisense™ and Motoman’s MotoSight™ 3D. The same is true for force sensors that will be required for fine assembly or insertion tasks.
Finally, the tooling that will interact with the parts will need to accommodate the variability. Universal grippers such as the Robotiq Adaptive Gripper will enable picking different parts using a single tool in assembly tasks.
Industrial robots turned 50 this year. In the next few years, we might see a new branch in their evolution, removing barriers between them and their human co-workers. This added flexibility should bring exciting opportunities to all levels in the robotic industry. For the end-user in particular, having a robotic system that can evolve with their production needs while taking full advantage of their skilled workforce will reassure them that the investment in equipment they make today will still be useful in five years. If these devices become widely accepted by the automotive industry, their affordability should become sufficient to be used by both large and small companies with the need to automate customized production runs.
Robotiq, a member of Robotic Industries Association, designs and manufactures flexible robotic grippers, working with robot manufacturers and integrators to deliver dexterous grippers that enable new applications and reduce time-to-market. The robotic components manufacturer is based in Quebec City, Canada and is an RIA Supplier Member. For further information, contact (888) 762-6847, or visit www.robotiq.com.