TL;DR
This paper introduces "zoomorphic gestures" – inspired by dog body language – as a novel way for industrial robots (cobots) to communicate their status to human collaborators. Through user studies, they demonstrate that these gestures are more intuitive and preferred by users compared to abstract gestures or simple light displays, offering a more engaging and understandable human-robot interaction.
What Problem Does This Paper Address?
Effective communication between humans and collaborative robots (cobots) is crucial for safe and efficient industrial operations. While human-to-robot communication is well-studied, robot-to-human (RtH) communication, especially conveying the robot's state, has received less attention. Existing methods like acoustic signals can be drowned out by industrial noise, and mechanical feedback has limitations. Visual communication, such as lights or text, can lead to information overload or require specific robot designs. Gestures, particularly those based on human body language, are intuitive but can create unrealistic expectations for non-humanoid cobots or lead to a perceptual mismatch. This paper addresses the need for a clear, intuitive, and engaging RtH communication method for appearance-constrained industrial cobots, proposing zoomorphic gestures as a solution.
Methods
The researchers developed three visual communication modalities for specific industrial use cases (e.g., greeting, error reporting):
- Zoomorphic Gestures: Inspired by dog body language, these gestures mimic canine behaviors to convey robot intentions. They were designed to be "dog-like" by exploiting structural similarities (e.g., camera as eyes) and ensuring a natural flow.
- Abstract Gestures: These are novel, robot-specific movements not based on any existing body language. They were designed to be simple, goal-oriented, and efficient, with a more "robot-typical" movement style.
- Light Displays: This served as a baseline modality, using color and blinking frequency to indicate robot status, based on established guidelines for electronic devices.
Two user studies were conducted. The first study (n=93) aimed to identify the best design option within each modality and evaluate the effectiveness of the design guidelines. The second, larger study (n=214), served as a benchmark to compare the three modalities (using the best options identified in study I) in terms of intuitive understanding, user experience, and user preference.
Datasets
Not discussed in the paper. The researchers created their own stimuli (videos of gestures and light displays) for the user studies.
Key Insights and Ideas
- Dogs as a Communication Bridge: The inherent social dynamic and familiarity humans have with dog body language makes it a surprisingly effective and intuitive communication channel for robots, even for those without prior dog ownership experience.
- Beyond Human Mimicry: For non-humanoid robots, mimicking human gestures can be problematic, leading to uncanny feelings or mismatched expectations. Zoomorphic gestures offer a way to leverage the intuitiveness of gestures without these drawbacks.
- "Dog-like" Design Principles: Exploiting structural similarities between the robot and a dog (e.g., camera as eyes) and ensuring natural, flowing movements are key to creating effective zoomorphic gestures.
- User Preference for Movement: Despite the potential for abstract gestures to be efficient, users overwhelmingly preferred movement-based communication (both zoomorphic and abstract) over static light displays.
Key Results
- Intuitive Understanding: In the benchmark study, zoomorphic and abstract gestures achieved significantly higher classification accuracy (around 58%) compared to light displays (around 39%).
- User Experience: Zoomorphic gestures were rated significantly higher across attractiveness, joy of use, and intuitiveness compared to abstract gestures and light displays.
- User Preference: A strong majority (56.1%) of participants preferred zoomorphic gestures for communicating robot states, citing reasons like "logical," "human," and "likable." Abstract gestures were preferred by 21.1%, and light displays by 16.8%.
- Perception Differences: Zoomorphic gestures were rated higher for anthropomorphism and animacy, while light displays were perceived as safer, though the effect size for safety was small.
Limitations and Future Work
- Safety Considerations: The current work did not explicitly address safety aspects, which are critical for movement-based communication in industrial settings. Future work should incorporate safety.
- Contextual Information: The videos used for light displays had less contextual information than those for gestures, potentially influencing results. Future studies should ensure equal context.
- Gesture Duration: More experienced users might find the longer execution times of gestures to be a hindrance in fast-paced industrial environments.
- Exploration of Modalities: The study explored a limited design space for light displays and gestures. Further research with different design options and use cases is needed to generalize findings.
- Certainty of Classification: Future studies could benefit from asking participants about their certainty in classifying the robot's state.
- Additional Use Cases and Modalities: Developing gestures for more complex use cases and exploring combinations of visual and non-visual modalities are suggested future directions.
Summary updated