What are cobots and how will they impact the future of manufacturing?
In the era of automation and robotization, the way humans work together with robots is becoming ever more important. In our latest blog post, we look at the development of industrial cobots (collaborative robots), and how they’ll impact whole industries in the not too distant future.
Across the spectrum of smart technology, it’s clear that everything from computers and cars to heavy machinery, and MedTech are becoming more and more intelligent every day. With new artificial intelligence algorithms, the technology around us is constantly adjusting to our needs and behaviors to serve us in the best way. This technology assistance is also scaling up. Industrial ‘cobots’ or collaborative robots are designed to be able to collaborate with humans in an intelligent and safe manner, and are set to become a key part of industry 4.0.
I have always been fascinated by technology that mimics human intelligence. It reminds me of what evolution does in nature – another subject I’m interested in. Evolution has led us to where we are today, and has shown us what works and what doesn’t. For humans, the knowledge transfer of a task, for example, from one person to another, usually results in learning best way of performing that task. The rest is unconscious know-how that’s not easily transferable.
However, cognitive science methods have come long way in capturing the know-how of humans and programming it into robots. In addition, the advancements in robotics, and digitalization are forcing us to formalize this knowledge, as well as the rules and regulations – both spoken and unspoken.
What are cobots?
Collaborative robots are robots that are capable of collaborating with humans. This collaboration is supposed to enhance human abilities in a safe way. In comparison, robotics deployments that do not assume human-robot collaboration typically function independently from humans, and often reside in a cage. They can also be programmed to stop when a human enters the facility where the robot operate. This, in turn leads to unwanted delays in operation or production, which can be avoided through the use of cobots. Collaborative robots are capable of monitoring the environment and co-existing in the same facility together with humans without sacrificing performance or safety.
The industry development of cobots is ongoing in several different areas. Faster reaction time, more exact movement patterns, orientation capabilities, capabilities in imitating humans – all these aspects contribute to advancements in cobot development. In addition, brain-computer interfaces is an exciting area that has made significant progress recently. When brain signals can be read with high precision and transferred to the robots, we will be able to collaborate with them in a completely new way.
Where cobots will offer the most value
Cobots offer the most value in situations where a human needs to be in close proximity to the robots. This includes human-robot collaborations where a human guides the robots, monitors the process, or even learns from the robots.
There is a clear fit when it comes to cobots and 5G systems. One of the key characteristics of cobots will be quality of service requirements, which vary over time and context – smart manufacturing for instance. In this recent paper, we describe an example of cobots inside a manufacturing facility, with safety zones associated with each cobot. The safety zones vary depending on the proximity of obstacles, proximity of humans in the same facility, and the speed of movement of the cobot. For the control system to be able to react and take timely decisions, for example, stopping the movement of a cobot, efficient and reliable communication is highly important.
A cobot in a facility without humans does not need as much attention as a cobot in close proximity to a human. Hence, the need for a high amount of signaling, high bandwidth, low latency, and fast decision-making capabilities through efficient computing is higher for the cobot in more safety-critical environments. 5G offers efficient network slice allocation, providing the necessary quality of service, which optimizes resource utilization in mobile networks as a result.
There is a lot of ongoing research in to how robots can mimic human movements and decisions. On the other hand, artificial intelligence algorithms have proven to be capable of finding solutions that are superior to those discovered by humans. We’ve seen examples of this in game strategies and design optimization problems. I think it would be very interesting to use cobots to make human work more efficient, and the collaboration between humans and robots more efficient.
What will be the impact of cobot deployment?
Reliability, safety and trust. The moment a technology becomes trustworthy it also becomes widespread. When people don’t trust their device or application, they put tape on microphones or their laptop cameras, for instance. When vehicle manufacturers don’t trust the quality of the mobile network, they don’t execute anything critical on the cloud, only on the vehicle. Similarly, if we don’t trust industrial robots, we would be reluctant to collaborate with them. This is one of the main challenges in cobot development.
In addition, cobots take many decisions in real-time, and the search space can be very large. This means traversals of large knowledge graphs, which, in turn requires processing power, a reliable network and novel software architectures. However, recent developments in technologies such as linked data, parallel processing, edge computing and distributed artificial intelligence allow for efficient decision making by cobots, making execution robust and efficient.
The challenges of deploying cobots at scale
Massive deployments of any new technology is a chicken-and-egg problem, because industries need to develop use cases that make use of cobots. Historically, all automation using robotics has been configured and tuned in the design phase. In a best case scenario, it has been adjusted at run-time in a data-driven approach. Here, cobots give process developers the possibility to teach cobots how to perform tasks, and actively evolve at run-time, allowing for improvements through artificial intelligence. The technology is there, but the use cases need to be defined at scale in order to achieve massive deployments at scale.
A challenge with the market deployment of cobots is that insufficient technology maturity hinders the market deployment of cobots. Cobot technology includes hardware design, sensors and actuators, efficient information processing, video processing, planning and multiple of fields from artificial intelligence landscapes, along with technologies that ensure safety, predictability and security of the solution.
What are the next steps for:
- service providers?
Service providers will need to be ready to provide reliable, differentiated quality of service through their networks. Safety and predictability requirements of applications (cobots, in this case) propagate requirements on the network infrastructure, so service providers will need to guarantee a certain quality of service for safety-critical situations.
- technology vendors?
Technology vendors will need to ensure trustworthiness of the systems, in terms of predictability, security and safety. They’ll also need to ensure adherence to regulations and compliance to standards.
Enterprises will need to develop use cases that benefit from human-robot collaboration, as this technology matures.
As evolution shows, we must adapt if we’re to survive in the long term. Human-robot collaboration is on the horizon and everyone from manufacturers to service providers has an opportunity to reach in to this market. Precision accuracy, improved efficiency, versatile operations, and reduced human-error risk factors, all mean that we’ll be seeing much more of this technology in the near future.
Find out how we’re unleashing the power of remote and cloud robotics.
Read our blog post, ‘The future of manufacturing is smart, secure and stable’
Explore how we’re driving 5G for manufacturing