A recent virtual reality (VR) study explored the impact of autonomous prosthetic arm movement speed on user perception, usability, and the sense of body integration.
The research found that a movement duration of approximately one second consistently resulted in higher ratings for body ownership, agency, and usability compared to significantly faster or slower speeds.
These findings suggest that the design of future autonomous prosthetics should prioritize user acceptance and integration over maximizing speed alone.
Study Overview and Methodology
This study utilized virtual reality to investigate how the movement speed of an autonomous prosthetic arm influences user embodiment (sense of body ownership and agency), usability, and social impressions of the robotic limb (competence, warmth, and discomfort). The research aims to address potential challenges in user acceptance as prosthetic technologies increasingly incorporate autonomous movements.
Nineteen male university students participated in a VR simulation where their virtual avatar's left forearm was replaced by a prosthetic limb. Participants wore a high-resolution head-mounted display and a motion capture suit, with a rigid brace on their real left arm to isolate the prosthetic's autonomous movement.
The core task involved reaching for a purple sphere. Once the participant's upper arm brought the virtual elbow close to the target, the virtual prosthetic forearm autonomously flexed towards the sphere.
Researchers systematically varied the autonomous movement duration across six distinct levels:
- 125 milliseconds (fastest)
- 250 milliseconds
- 500 milliseconds
- 1 second
- 2 seconds
- 4 seconds (slowest)
After each block of 15 reaches, participants completed questionnaires to assess their sense of body ownership, sense of agency, usability (using the System Usability Scale - SUS), and social impressions of the robot (using the Robotic Social Attributes Scale - RoSAS, which measures competence, warmth, and discomfort).
Key Findings
The study identified a movement duration of approximately one second as optimal across several metrics, with both extremely fast and slow movements yielding less favorable results.
Embodiment and Usability
- A moderate movement speed, specifically a duration of approximately one second, resulted in the highest ratings for body ownership, sense of agency, and usability.
- The 500 ms duration also performed significantly better than the speed extremes.
- Conversely, both excessively fast (125 ms) and excessively slow (4 s) movements significantly reduced body ownership, sense of agency, and usability scores.
Social Impressions
- Competence: Perceived competence of the prosthetic arm was higher at moderate to slightly faster speeds (500 ms and 1 second) compared to the slowest speed (4 seconds), and at 1 second compared to 2 seconds.
- Discomfort: The fastest movement (125 ms) generated significantly higher discomfort ratings compared to all other tested conditions (500 ms, 1 s, 2 s, 4 s).
- Warmth: No significant correlation or dependence on movement speed was observed for warmth ratings.
The researchers suggested that the optimal 1-second duration aligns with the expected timing of natural human reaching movements when accuracy is prioritized.
Additional Observations and Implications
An additional behavioral finding was that participants adjusted their own upper arm movement speed to match the prosthetic's expected autonomous speed, slowing their movements when the prosthetic was set to move slowly.
The findings suggest that for future AI-enabled prostheses offering autonomous assistance, simply aiming for faster and more accurate performance may be insufficient. Instead, the design of movement speed should prioritize matching what users can readily accept and integrate as a natural extension of their own body. These insights may also extend to the design of other robotic body augmentation technologies, such as supernumerary robotic limbs, exoskeletons, and wearable robots.
Study Details and Future Research
The study acknowledged limitations due to its virtual nature and the use of healthy participants. It did not account for real-world factors such as the physical forces, weight, or connection point forces of a physical prosthesis. Additionally, the constant visibility of the target made the prosthetic's intention highly predictable, which could influence agency and ownership.
Despite these limitations, virtual reality proved to be a valuable tool for safely simulating and evaluating advanced prosthetic technologies.
Future Research Directions
- Exploring long-term user adaptation and learning, investigating if continuous use of fast and accurate robotic body parts can lead to increased embodiment and a sense of normality over time.
- Investigating less predictable scenarios.
- Incorporating physiological measures.
The research was supported by JSPS KAKENHI, the Murata Science and Education Foundation, JST, and MEXT, and published in the journal Scientific Reports.