GelForce is a novel interface that measures the distribution of both the magnitude and direction of force. The sensor is composed of a transparent elastic body, two layers of blue and red markers, and a CCD camera. Force vectors are calculated from the captured movement of the markers.
The GelForce sensor measures information that cannot be obtained by existing sensors, so it is a promising tool for a myriad of applications. It can be used as an intuitive computer interface like the touchpad or mouse, but with a much richer geometrical structure. This characteristic makes it a powerful way to interact with virtual environments modeled with physical parameters. For example, it could be used as a molding instrument for realistic sculpting and image processing. It also provides lush external input for real-time physical simulations.
The sensor is also an adept way to endow robotic and prosthetic devices with tactile sensations similar to human capabilities. When we pick up a fragile object such as a glass, we perceive its weight as a force parallel to our skin, and we respond by controlling the grasping force to prevent slippage. When the moldable GelForce sensor is combined with a robotic hand, the robot can receive force information parallel to the fingertip and respond with the same degree of dexterity.
It would not be an exaggeration to say that every scientific technology is based on sensors. For example, audio recording cannot exist without measurement of sound, and television broadcasts would be impossible without converting radio waves to image data. The GelForce sensor could become equally essential. It can measure all information about force: direction, magnitude, and distribution. And its sensing mechanism is quite simple, so the sensor can adopt various shapes. Popular devices that incorporate tactile sensation are certain to become mainstream in the near future.
One of the goals of this project is to develop a finger-shaped sensor and attach it to a robotic hand so that robots with tactile sensation can perform fine manipulation tasks just as humans do. Existing sensors on mechanical hands cannot measure sufficient force information to perform most everyday tasks, which severely limits the development of field robotics. Developing a sensor with a finger shape confronts two key technical issues:
1. The sensor must be a small high-density unit, comparable to the layout of mechano-receptors in the skin.
2. The force sensor must adapt to the particular form of the fingertip.
The simple structure of this elastic sensor enables an elegant solution to these two issues. The sensor does not require a complex sensing unit. It uses markers within a gelatin body, which can be arranged at high density. In addition, because the sensor only needs to capture the markers, it can be crafted into any shape that allows it to do so.
The core innovation is the technique to derive distribution-of-force vectors from the movement of two layers of colored markers. If the sensor includes only one layer of markers, only the magnitude of force can be measured with any degree of accuracy. Two cameras (for stereoscopic vision) would be required to measure the three-dimensional movement of markers. But using two cameras would require complex camera calibration and a much larger system. These complications would make downsizing the sensor infeasible. The GelForce sensor solves this problem with two layers of markers at different depths, which delivers sufficient measurement information to readily obtain a force-vector distribution.
Tuesday, 10 August
10:30 am - 12:15 pm
The University of Tokyo
The University of Tokyo