17:15 〜 19:15
[AOS14-P02] Evaluating the Effect of Gravitational Force Sensing on Visual Gravitational Motion Prediction
キーワード：重力、ヒトの運動、運動予測、VR
Astronauts on long-duration space missions are exposed to zero-to-microgravity, which leads to muscle atrophy and immune dysfunction. Current countermeasures, such as daily exercise, have proven insufficient in fully mitigating these effects. Artificial gravity, generated through centrifugal force, has been proposed as a potential solution. However, it introduces varying gravitational loads across different body regions, which may impact motor prediction. The vestibular system, responsible for detecting gravity via the otolith organs, plays a critical role in balance and posture control. In unfamiliar gravitational environments, discrepancies in motor prediction may contribute to space motion sickness and reaction errors. Despite this, few studies have quantitatively evaluated these effects.
This study investigates the impact of visual-gravitational mismatches on motion prediction. Participants were either seated or lying supine while wearing a head-mounted display (HMD) that presented virtual reality (VR) simulations created using Unity and LabVIEW. A 10 cm pink ball was shown falling from a height of 5 m, and participants were asked to predict the time of its arrival at a 2 m target. The experiment examined five initial velocities (0.0, 1.0, 1.5, 2.0, and 4.15 m/s²) under two gravity conditions (1G and 1/6G) with varying visual exposure durations.
• Experiment 1: 20 trials under 1G with no feedback.
• Experiment 2: 20 trials under 1G with feedback, 10 trials under 1/6G (seated, with feedback), 20 trials under 1/6G (both postures, no feedback), 10 trials under 1/6G (seated, with feedback), and 10 trials under 1G (seated, with feedback).
• The order of trials was randomized for each participant.
Prediction errors were calculated using Time-to-Contact (TTC) and analyzed using MATLAB and JASP. In seated trials, an increase in initial velocity resulted in reduced prediction errors. In supine trials, errors were generally larger, particularly with short visual exposure durations. Under 1G conditions, errors remained near zero due to prior learning, while under 1/6G, errors initially deviated negatively but gradually approached zero. During the generalization phase in 1/6G, seated participants exhibited increasing errors with higher velocities, whereas supine participants showed smaller overall errors.
Statistical analysis revealed no significant differences in prediction errors based on posture, exposure duration, initial velocity, or visual gravitational acceleration. Additionally, a quantitative evaluation was conducted to determine the gravitational values participants relied upon for estimation. Future research should explore long-term adaptation mechanisms to optimize artificial gravity training for space missions.
This study investigates the impact of visual-gravitational mismatches on motion prediction. Participants were either seated or lying supine while wearing a head-mounted display (HMD) that presented virtual reality (VR) simulations created using Unity and LabVIEW. A 10 cm pink ball was shown falling from a height of 5 m, and participants were asked to predict the time of its arrival at a 2 m target. The experiment examined five initial velocities (0.0, 1.0, 1.5, 2.0, and 4.15 m/s²) under two gravity conditions (1G and 1/6G) with varying visual exposure durations.
• Experiment 1: 20 trials under 1G with no feedback.
• Experiment 2: 20 trials under 1G with feedback, 10 trials under 1/6G (seated, with feedback), 20 trials under 1/6G (both postures, no feedback), 10 trials under 1/6G (seated, with feedback), and 10 trials under 1G (seated, with feedback).
• The order of trials was randomized for each participant.
Prediction errors were calculated using Time-to-Contact (TTC) and analyzed using MATLAB and JASP. In seated trials, an increase in initial velocity resulted in reduced prediction errors. In supine trials, errors were generally larger, particularly with short visual exposure durations. Under 1G conditions, errors remained near zero due to prior learning, while under 1/6G, errors initially deviated negatively but gradually approached zero. During the generalization phase in 1/6G, seated participants exhibited increasing errors with higher velocities, whereas supine participants showed smaller overall errors.
Statistical analysis revealed no significant differences in prediction errors based on posture, exposure duration, initial velocity, or visual gravitational acceleration. Additionally, a quantitative evaluation was conducted to determine the gravitational values participants relied upon for estimation. Future research should explore long-term adaptation mechanisms to optimize artificial gravity training for space missions.