A Concept for User-Centered Delegation of Abstract High-Level Tasks to Cobots for Flexible Lot Sizes

Overview

• Collaborative robots (cobots) are becoming more popular in companies, but human operators are often not trained to program complex machines.
• Humans typically communicate on a task-based level rather than through specific instructions, which can lead to low usability scores for teach pendant programming.
• To address this, the researchers propose a task-based interaction concept that allows operators to delegate complex tasks to a cobot without programming, by specifying tasks using "triplets."
• The concept involves task decomposition and a reasoning system using a cognitive architecture.
• The approach is evaluated in an industrial use case where a cobot is used to sand mineral cast basins in a crafts enterprise.

Plain English Explanation

Collaborative robots, or cobots, are becoming more and more common in companies. These robots are designed to work alongside human workers, rather than being fully automated. However, many human operators are not trained to program these complex machines.

Humans are used to communicating with each other on a higher-level, task-based level rather than using the specific, low-level instructions that machines typically require. This gap between how humans and machines communicate can lead to cobots being difficult for people to use and program.

To solve this problem, the researchers in this paper propose a new way for humans to interact with cobots. Instead of programming the cobot directly, the human operator can simply describe the task they want the cobot to perform, using a special set of "triplets." The cobot can then use this task description, along with its own reasoning and cognitive architecture, to figure out how to actually carry out the task.

The researchers tested this approach in an industrial setting, where a cobot was used to sand mineral cast basins in a crafts workshop. This allowed the human workers to delegate this complex task to the cobot without having to directly program it.

Technical Explanation

The key innovation in this paper is a task-based interaction concept that allows human operators to specify complex tasks for a cobot to perform, without requiring the operator to program the cobot directly.

The concept is based on task decomposition and a reasoning system using a cognitive architecture. The human operator specifies the task they want the cobot to perform using a set of "triplets" - a high-level description of the task goals, constraints, and context.

The cobot's reasoning system then takes this task description and uses it to plan and execute the necessary actions to carry out the task, without the human operator having to provide detailed, low-level instructions.

The researchers evaluated this approach in an industrial use case, where a cobot was used to sand mineral cast basins in a crafts enterprise. This allowed the human workers to delegate this complex task to the cobot without having to directly program it.

Critical Analysis

The researchers note that this task-based interaction concept is a promising approach for improving the usability of cobots, but acknowledge that there are still some limitations and areas for further research.

For example, the current implementation relies on a pre-defined set of task "triplets" that the human operator must choose from. It would be valuable to explore ways to allow for more flexible, natural language-based task descriptions. Additionally, the cognitive architecture used by the cobot could potentially be expanded to handle a wider range of task types and contexts.

Another potential concern is the robustness of the task reasoning and planning algorithms. In real-world industrial settings, there may be unexpected events or variations that the cobot needs to be able to handle gracefully. Further research is needed to ensure the reliability and adaptability of this approach in more complex, dynamic environments.

Overall, this research represents an important step forward in enabling more natural, intuitive collaboration between humans and cobots. By bridging the gap between task-based human communication and instruction-based machine interaction, it has the potential to significantly improve the usability and adoption of collaborative robots in industry.

Conclusion

This paper presents a novel task-based interaction concept that allows human operators to delegate complex tasks to collaborative robots (cobots) without the need for direct programming. By using a task decomposition and reasoning system based on a cognitive architecture, the cobot can understand and execute tasks specified at a high level by the human operator.

The researchers demonstrated the effectiveness of this approach in an industrial use case involving a cobot sanding mineral cast basins in a crafts enterprise. This suggests that this task-based interaction concept could help address the usability challenges that often limit the adoption of cobots, by enabling more natural, intuitive collaboration between humans and machines.

While there are still some limitations and areas for further research, this work represents an important step forward in the field of human-robot collaboration and flexible robot task planning. As collaborative robots become more prevalent in industries, innovations like this will be essential for unlocking their full potential to work seamlessly alongside human workers.