Thanks for visiting Imaging and Machine Vision Europe.

You're trying to access an editorial feature that is only available to logged in, registered users of Imaging and Machine Vision Europe. Registering is completely free, so why not sign up with us?

By registering, as well as being able to browse all content on the site without further interruption, you'll also have the option to receive our magazine (multiple times a year) and our email newsletters.

Collaborative co-workers

Share this on social media:

Topic tags: 

The nature of human and robot collaborations within factories is evolving. Beth Harlen investigates the influence of machine vision technology on this budding relationship

Since the first industrial robot was put to work in a General Motors automobile factory in 1961, manufacturers have sought to improve efficiencies by harnessing the potential of these machines. Expensive safety fences surrounding operational robots have long been a necessity for protecting human workers, but technologies are being researched and developed with the aim of ensuring a true collaborative relationship between humans and robots. Experts within the industry, such as Fraunhofer IPA’s Milad Geravand, believe that the introduction of human-robot collaboration (HRC) will have a dramatic impact on the factory floor. 

‘The most noticeable change will be that safety fences will no longer be required with HRC systems, which will allow the building of robot systems with a smaller footprint and shorter distances for material handling,’ Geravand explained. ‘The resulting systems will exceed the agility and flexibility of today’s robots by far and the inclusion of the robot in the process will allow us to respond to changes and problems much quicker. This [also] promises to address the demographic change in Western societies, by improving working conditions and physical requirements for workers. While robotics had the reputation of a job killer 10 years ago, this development has made robotics a key technology by keeping jobs in Europe in the face of an ageing society.’

A further benefit, according to Mathieu Bélanger-Barrette, production engineer at Robotiq, is that with the lack of a fully qualified workforce, manufacturers need to find a way to remove qualified workers from redundant tasks and repurpose them to jobs that need more dexterity or decision. ‘Another great aspect resulting from collaborative robots is the fact that humans can work safely and do tasks that are not monotonous,’ he commented. ‘In fact, repetitive tasks can be harmful for a human when repeated several thousand times. By leaving the robot doing this redundant task, the worker can do qualified jobs and feel more engaged in the work.’

Bélanger-Barrette added that SMEs and global companies can benefit from these robots because of their fast return on investment and ease of programming. Because no additional guards are required for these robots, the price of the robot cell is reduced. In addition, no special expertise is needed to program them, saving further costs.

Cause for concern?

While on the surface HRC seems to be the perfect solution for modern factories, it is not without its concerns. Earl Yardley, director at Industrial Vision Systems, emphasised the absolute priority of worker safety in all robotic environments. ‘Measures must be introduced to avoid robot-related accidents when humans and robots are sharing the same workspace,’ he said. ‘This can include using simulation software to plan out and test robotic concepts, create adaptive zones, define robot speeds, and revisit the regulations. The majority of manufacturers are concerned about how to ensure employee safety when working alongside collaborative robots, so risk assessments still need to be completed and, in some situations, collaborative robots will still need to be guided.’

Geravand added that the big challenge is the ‘robustness’ and ‘reliability’ of the vision systems required for these industrial applications. For example, if a vision system is going to be used for the purpose of safety in human-robot collaboration, this system should give correct data under any lighting conditions. He is of the opinion that machine vision can indeed supply the necessary information to avoid hazards and guarantee human safety when working with robots. ‘For productivity and cost effectiveness, machine vision can provide all the required information that allows a truly productive interaction,’ Geravand commented. ‘The improvements in sensor technology have resulted in quantum leaps in machine vision performance in recent years, and the basic principle of any collaboration with a robot is to avoid any hazard to the human, such as through collisions. 

‘This can be achieved by a proper vision system that reports the existence of humans next to the robot. However, to achieve this, the vision system needs to have a certified safety integrity ensuring that the human can be detected reliably.’ He added that a robot system can perform more sophisticated tasks when working alongside humans when it’s empowered with a good monitoring system.

Focusing on the aspect of safety, the Schmersal Group has recently sponsored the BeyondSPAI research project at the Bonn-Rhein-Sieg University of Applied Sciences, with funds totalling €45,000. André Batz, director of research and development at the Schmersal Group, commented: ‘The aim of the research project is to use camera systems to develop solution strategies that allow the robot to identify people reliably, and to differentiate between humans and objects or workpieces. The image information is then used to derive strategies for hazard prevention. The robot controller could, for example, trigger a robot shutdown, or alternatively reduce the speed of movement or divert the robot movement to allow humans and robots to work in close physical proximity. However, this requires an absolute reliability in optical recognition of humans.’

A few years ago, Schmersal developed the safety controller, a safety solution for human and robot collaboration that specifies a definable three-dimensional working zone for the robot, registering and shutting it down immediately if it moves outside the area. Scientists working on the project at the University of Bonn-Rhein-Sieg are now investigating how camera systems and special image processing algorithms can be used to allow robots not only to identify human silhouettes, but also to recognise human skin. If successful, this development should enable industrial robots to register if a person comes too close, in order to shut down operations in good time. This would mean robots could leave their virtual working area and so be used more flexibly.

Director of the project, Professor Dr Norbert Jung of the Institute for Safety Research, emphasised that vision-based approaches offering the possibility of detecting entire silhouettes of people and different postures is very important for collaboration. This is because the position of every part of the human – especially their limbs – has to be known by the protective device to make the environment safe in the comparably small zone surrounding the robot. ‘It can be expected that [machine vision] will play an important role here,’ he said, adding that skin detection is based on the SWIR wavelength range. ‘We expect that machine and computer vision still has to be supplemented by additional measures to achieve the targeted reliability in person detection,’ Jung commented. 

Enabling technologies

When it comes to facilitating the collaborative relationship between humans and robots, the most suitable machine vision technology strongly depends on the targeted application and its requirements. Fraunhofer IPA’s Geravand explained that for lightweight robots there have been many attempts to use available sensors, such as cameras, lasers, Microsoft Kinect devices, and so on, to enhance their performance when working with humans. Lightweight robots are safer because of both their low mass and the addition of sensors that detect when the robot comes into contact with humans.

These sensing approaches can also be used for heavy payload robots. However, here a non-contact safety principle is required as these robots store significant amounts of energy. ‘The use of machine vision in these applications is very intriguing, but the requirements on safety integrity come to mind. Here, vision based solutions like the Safety Eye by Pilz are finding their way into industrial applications,’ said Geravand.

Robotiq’s Bélanger-Barrette believes that 3D vision systems are really promising, in terms of both safety and part recognition. However, for the moment he feels it’s difficult to implement a reliable 3D sensor, and that the focus is still on 2D vision – a proven technology with low cost and compact dimensions. Additionally, he said that the level of intelligence and image analysis will improve; it will be a lot faster and a lot more accurate than it currently is.

The ties between machine vision and robotics are certainly cementing, and Oxfordshire-based Industrial Vision Systems (IVS) is now integrating its machine vision technology with the latest collaborative robots, also known as ‘co-bots’, from Universal Robots. Supplied in the UK by RA Rodriguez (UK), the vision-enabled UR3, UR5 and UR10 models are aimed at applications such as inspecting complex components, as well as delivering positioning feedback to the robot.

Yardley at IVS commented: ‘We are creating vision systems for collaborative robots to enhance our overall productivity levels. This new generation of affordable lightweight robots is unlocking new markets and applications as they work side by side with human workers.’ He added that collaborative robots are eliminating the need for costly precision fixtures, meaning different parts can be processed and inspected without changing tools.

‘By using UR3, UR5, and UR10 model robots from Denmark’s Universal Robots – one of the earliest manufacturers of co-bots – our vision systems can be integrated with all major robot manufacturers and control systems,’ Yardley said. The vision system can therefore be used to inspect complex components, as well as giving the robot positioning feedback. The UR3 is a smaller tabletop robot that can be used to provide camera inspection on multiple sides of complex parts and components, allowing precision automated inspection.

Opening up opportunities

Beyond inspection, human-robot collaboration has an important role to play in industry. According to Bélanger-Barrette, machine tending – where robots replace human interaction with a production machine by transporting parts to and from the machine – is probably the sector where collaborative robots are used the most. ‘With a small footprint and a compact design, these robots can be fitted directly on a computer numerical control (CNC) machine with a very few additional devices,’ he explained. ‘That being said, the real advantage of these robots is that they can be repurposed on another machine or on other types of parts. In the reality of a machine shop, this is a huge advantage as you can run a small batch of parts and be up and running within a couple of minutes, as opposed to classical industrial robots that were designed for thousands of cycles and needed to be programmed using complex methods.’

The Schmersal Group’s Batz concluded by saying that, for Schmersal, there is no question that the demand for innovative safety technology for human-to-robot collaboration will increase further still. ‘While the number of industrial robots sold around the world was around 240,000 in 2015, a study by the International Federation of Robotics (IFR) predicts it will increase to 1.3 million by as early as 2018. And after South Korea and Japan, Germany has the highest density of robots around the world and is among the pioneers when it comes to automation with industrial robots.

‘If, in the future, robots work hand-in-hand with human specialists, rigid production processes can increasingly be replaced by flexible structures. This means robots can help increase efficiency in production. After all, robots have digital interfaces that allow them to integrate well into the networked structures of the “smart factory”. They are therefore a driving force behind the development of Industry 4.0 strategies.’