phonlamaiphoto -

Industries evaluate collaborative robot applications

Robotics has traditionally been viewed as a threat to jobs, but putting AI in robots can make machines more collaborative and empower human workers to be more effective.

Physical robots have been around for decades. For most of that time, they have been quarantined off from their human counterparts and supervisors, as they were too dangerous to work with humans side by side.

Then, in the mid-1990s, collaborative robot applications began to emerge. Known by its shorthand term cobot, a cobot is a physical robot that is intentionally designed to operate in close quarters with humans. Cobots can operate in a wide range of environments, from assembly lines to warehouses to roaming around hospitals or office buildings helping with various tasks.

Cobots tackle a wide range of tasks

While industrial robots have evolved to perform complicated tasks, they are only as good as what they are explicitly programmed to do. A robot that is programmed to pick boxes and pallets in a warehouse is not capable of welding doors on cars. This means that robots cannot be easily retrained to do another job, as there is usually significant cost and time involved to reprogram the robots.

Cobots, however, learn tasks via deep learning -- demonstration and reinforcement learning -- which makes them easier to train and capable to perform more than one task because they are not hardcoded. This also enables a more agile application of bots with frequently modified training and use by smaller companies that can purchase one bot and train it for multiple tasks.

In addition, as cobots become a bigger part of the extended workforce, the need for bots to interact with other cobots, as well as with their human counterparts, will become greater.

Cloud-connected cobots enable collaboration between two or more robots within the connected environment. This means that two bots that are not physically next to each other or even in the same geographic location can interact to accomplish a greater set of tasks. This connectivity combined with better data and information collection makes networked cobots significantly more valuable to enterprises.

Augmenting, not replacing, human work

The idea of artificial intelligence-enabled bots enhancing rather than replacing human workers is something that's top-of-mind for most enterprises. Not only are companies faced with resistance from human workers who need to train the bots, but it becomes a socioeconomic issue when automation replaces large portions of the workforce.

To this end, one of the most important issues with collaborative robot applications is using them to augment rather than replace human activity. Humans have creativity, common sense and intuition. Machines are great at probabilistic thinking, handling large volumes of data and being trained. When you combine these forces together, you give workers superhuman abilities.

As a result, many companies that have implemented cobots report greater adoption when cobots are used alongside humans than they do with robots that are meant to replace human activities. When the conversation changes and people are told cobots are meant to be collaborative tools rather than replacement tools, people feel less threatened and companies get more productive workers.

Early companies offering cobot technology, such as Kuka, Rethink Robotics and others, have focused on the augmented experience. Some of these vendors have included interactive features that are meant to assist their human counterparts and indicate simulated emotions. For example, a screen on Rethink Robotics cobots can indicate various levels of interest, frustration, happiness or boredom so that the human worker can gauge how to improve their overall efficiency and productivity.

In another case, surgical cobots working alongside surgeons are being used to perform complex medical procedures, such as removing gallbladders or stitching internal tissue. The cobot, working with a surgeon, can learn tasks through demonstration and reinforcement learning, enabling the cobot to take over some of the more repetitive or mundane tasks, such as initial incisions, or the more risky tasks that require extreme precision and focus.

Challenges with cobot adoption remain

To many in the robotics field, the difference between cobots and traditional robots is semantic rather than technical. But industry adopters would disagree. They see both the change of terminology, as well as the conceptual difference as significant.

Collaborative robot applications are designed to have lower overall power and greater sensitivity and awareness of their surroundings so that they can work in close proximity to humans. But this lower power does limit the application of cobots to tasks where a large amount of strength is not required.

In addition, the array of sensors that detect obstacles, human interference or other situational awareness issues can make the use of cobots in crowded spaces or in complex areas difficult. However, cobot companies have evolved their offerings to increase the physical capacity of their bots with multiple arms, graspers, control assemblies and more. Indeed, some cobots have multiple arms and multifunctional graspers that can switch out as needed to handle complex tasks.

As with all things, adoption and widespread acceptance of new technologies take time. People need to feel comfortable and unthreatened by a new tool in their work environment. If we change the narrative of the role of robots from industrial machinery meant to be isolated from human contact to collaborative bots that are meant to assist their human counterparts with an increasing array of capabilities, then cobots will be seen as augmentative tools to help perform work tasks more efficiently.

Dig Deeper on Enterprise applications of AI

Business Analytics
Data Management