Recent research has unveiled a significant pattern regarding the connection users feel with prosthetic arms. When these devices move either too quickly or too slowly, individuals reported a diminished sense of connection and usability. Conversely, when the prosthetic arm operated at a moderate pace, akin to natural human reaching--approximately one second for the motion--participants expressed a heightened sensation of the arm being an integral part of their body.
Advancements in Prosthetic Technology
For those who have lost a hand or arm, prosthetics are essential for daily activities. Much of the current research focuses on enhancing the responsiveness of these devices to the user's intentions. This often involves interpreting biosignals, such as electromyography (EMG) and electroencephalography (EEG), to facilitate movement.
Simultaneously, the rapid evolution of machine learning and AI is paving the way for prosthetics that can assist users by moving autonomously in specific scenarios. These semi-autonomous systems could predict user needs and offer support automatically. However, when a limb begins to move on its own, it may evoke feelings of discomfort or detachment, posing a challenge for widespread acceptance.
Investigating Movement Speed and User Experience
Previous studies indicate that individuals are more at ease with autonomous movements when they comprehend the underlying purpose. Building on this premise, Harin Manujaya Hapuarachchi and colleagues conducted research to determine the influence of movement speed on user acceptance.
In a virtual reality setup, participants interacted with an avatar whose left forearm was replaced by a prosthetic limb. They were tasked with reaching for a target while the virtual prosthetic arm moved independently. The researchers varied the duration of each movement across six speeds (from 125 ms to 4 s). After each trial, participants assessed how much the arm resembled their own, their sense of control, usability (using the System Usability Scale), and their perceptions of the robotic arm through a standardized scale evaluating competence, warmth, and discomfort.
The Optimal Movement Duration
The findings were both consistent and noteworthy:
- At a moderate speed (1-second movement duration), participants reported the highest levels of body ownership, agency, and usability.
- In contrast, the fastest (125 ms) and slowest (4 s) speeds resulted in significantly lower ratings.
- Users perceived the arm as more competent at moderate to slightly faster movements, while discomfort peaked at the highest speed. The perceived warmth of the arm did not show a clear connection to speed.
In essence, enhancing the speed of a prosthetic arm does not necessarily improve its effectiveness. Aligning its movement with the timing of natural human motion is crucial for fostering the feeling of ownership among users.
Future Directions for Prosthetic Design
These insights suggest that future AI-enhanced prosthetics should prioritize timing that aligns with human movement rather than focusing solely on speed. Designers will need to adjust movement patterns to meet the brain's expectations for a natural limb.
The implications of this research extend beyond prosthetic arms. Other technologies that act as extensions of the body, such as additional robotic limbs, exoskeletons, and wearable robots, could also gain from movements that replicate the natural rhythm of human motion.
Researchers are also interested in how prolonged usage alters perception. Users often begin to feel that frequently used tools become part of their body. With regular use, even advanced robotic limbs may start to feel "normal," easier to operate, and more integrated into the user's experience.
Virtual reality plays a pivotal role in this research, enabling scientists to assess new prosthetic technologies and control systems in a secure environment before they become widely accessible. This method allows for early evaluations of psychological responses, user acceptance, and design considerations during the development phase.
This research received support from JSPS KAKENHI (JP22KK0158), the Murata Science and Education Foundation, JST (JPMJFS121), and MEXT (202334Z302).