IntuiTeach: Faster Robot Teaching

March 2023

Small and medium-sized enterprises are still underutilizing collaborative robots. This puts their competitive edge over larger players at risk. One of the reasons that prevents SMEs from leveraging collaborative robots more extensively is an insufficient availability of trained cobot operators. This prevents them from flexibly deploying cobots and thereby reduces the value cobots can create for their businesses.

With IntuiTeach we have developed a system that allows users to intuitively program collaborative robots with hand gestures and by moving them into desired positions with both hands. This internationally patented system makes cobots accessible to novice users and speeds up programming of standard cobot workflows by 20%.

The advent of Cobots

Until fairly recently, automation has primarily been a big-company advantage. Several factors have now begun to level the playing field, including the emergence of collaborative robots, also called cobots. Cobots have introduced many of the same benefits to small and medium-size enterprises (SMEs) that traditional robots and automation offered large organizations. As safe, highly compact, and relatively intuitive solutions, cobots can be implemented quickly and cost-effectively.

Despite the many advantages collaborative robots offer to SMEs, their uptake remains still limited. The reasons therefore are multifold. Factors such as a lack of awareness and scepticism with which current workers approach cobots contribute. Another key factor wich limits uptake of cobots in SMEs is a lack of expertise regarding their operation.

Although, as compared to traditional robots, cobots are easier to integrate and use, (re-)deployments of currently available systems still requires technical experts for operation. With tight labour markets, many SMEs face challenges when it comes to attracting such talent which limits their ability to fully leverage the benefits cobots hold for them.

Intuitive Methods for Cobot Teaching

With the advent of cobots, more intuitive and effective methods for programing have been introduced. Advanced producers deliver their cobots with intuitive web applications and kinesthetic teaching (robot arm can be manually moved to a desired position which can be saved to the robots program for execution) modes. These tools lower the barrier to start experimenting with cobots and enable novice users to teach cobots simple routines. Furthermore, kinsethetic teaching methods can significantly speed-up the programing process.

Nonetheless, if standard cobot workflows are to be implemented, these tools still require certain levels of programing experience. Furthermore, the user experience is not optimal, as users are required to frequently switch between bringing the robot into the right position and confirming certain positions on either a tablet or a laptop. On the one hand this reduces the efficiency of programing and on the other hand tablets and laptops are prone to frequently breaking down in industrial environments.

Therefore, it can be concluded that currently available methods for cobot teaching are not yet developed far enough as to enable mass adoption.

IntuiTeach - A new approach to cobot teaching

In our research we are exploring ways to further reduce the technical expertise required for programing cobots. One of the outcomes of our research is IntuiTeach - an intuitive method for teaching collaborative robots, which speeds up the process of programming standard cobot workflows by 20% (shown in a study conducted with 25 novice users).

Our technology enables the programming of collaborative robots without the need for peripheral devices such as laptops or tablets. Instead, the operator can program the cobot through intuitive hand gestures alone.

To do so, the operator wears an electromyographic armband on the underarm. This armband measures muscular activity based on which a machine learning algorithm identifies which hand gesture the operator is executing. Therefore, as soon as the operator grabs the cobot, our system identifies the “grabbing” hand position and activates the cobot’s free drive mode. This allows the operator to move the cobot into a desired position. With other hand gestures such as bending the hand outwards or making a fist, the operator can activate the end effector and save a new waypoint to the program for later execution by the cobot. (see video below)

Our experiments with 25 novice users have shown that this technology increases the speed of programing a standard cobot workflow by 20% (as compared to programing the same workflow using a publicly available state of the art cobot teaching method). Furthermore, participants reported that our teaching method was more user friendly. These findings were also confirmed in expert interviews.

Looking forwad

Looking forward, we believe our technology, especially in combination with advances in other areas, has significant potential to augment how cobots are taught. We see potential not only for novice and untrained but also for expert users as the technology comes with many operational benefits. We are particularly hopeful, that our technology will support SMEs in leveraging the advantages collaborative robots can bring to their business.