Assistive robots give humans a leg (and arm) up
Reading subtle muscle movements to anticipate what happens next
OSAKA -- Factory workers could carry multiple items by themselves more easily, and seniors with weaker legs could get a helping hand in getting out of chairs, thanks to Japanese robotics technology that predicts their next moves.
Sensing changes in muscles, the assistive devices exploit the unconscious movements that humans make when they think of moving a body part.
A team led by engineering professor Taro Maeda of Osaka University has developed a robotic arm that attaches to the hip. The person initially watches how it swings by itself or points at buttons.
When moving an arm, people predict how much force will be placed on the body and unconsciously alter their stance. They do so even if the movement is by a robotic arm extending from the hip. Changes in muscles will be captured by sensors attached to the shoulders and knees, then learned by the control system.
In actual use, the shoulder and knee sensors will sense changes in muscles to read how the person intends to move the arm. The robotic arm will then be moved. In testing that involved pointing at on-screen buttons to music, people performed the task at the right time with a 70%-plus success rate.
Panasonic has developed a chair that pushes the seat up when a person is about to get up. Changes in the posture of a seated person about to stand up were studied, along with the timing of pressure on leg muscles. Analysis showed that the order in which shin and thigh muscles move indicates that a seated person is about to stand.
How the user bends the upper body is measured by an accelerometer sensor attached to the hip, while muscle movements are gauged via wireless electrode. Tests were 99.5% successful in predicting that a seated person was about to get up.