Remember me - user-centered implementation of working memory architectures on an industrial robot

Area of Science:

• Human-Robot Interaction (HRI)
• Robotics
• Artificial Intelligence (AI)

Background:

• User-centered design is crucial for effective Human-Robot Interaction (HRI) but is underutilized in industrial robotics.
• Implementing advanced AI architectures like GRU and WorkMATe requires understanding user perceptions.
• Evaluating robot appearance, movement, and memory functions from a user perspective is essential for adoption.

Purpose of the Study:

• To apply a user-centered approach for implementing and training two working memory architectures (GRU and WorkMATe) on an industrial robot.
• To assess user perceptions of the robot's appearance, movements, and memory functions before and after AI implementation.
• To explore user ideas on robot memory and navigation, and identify factors influencing HRI.

Main Methods:

• Two online user studies with large, diverse samples were conducted.
• User Study 1 evaluated the robot before working memory implementation, gathering feedback on appearance, movement, and memory expectations.
• User Study 2 assessed user perceptions after implementing GRU and WorkMATe architectures, analyzing the impact on robot evaluation.

Main Results:

• User Study 1 showed moderate robot evaluations, influenced by participant age and negative attitudes towards robots, linked to low robot experience.
• User Study 2 indicated that implementing WorkMATe and GRU improved perceived robot memory but not appearance or movement.
• Positive user views on robots and their experience significantly drove overall robot evaluations.

Conclusions:

• User-centered design significantly enhances perceived robot navigation efficiency and user acceptance.
• Implementing working memory architectures positively impacts user perception of robot memory capabilities.
• User experience with robots is a critical factor for fostering positive and effective human-robot interaction in industrial settings.