MIT researchers have developed a new thin robotic finger designed to detect buried articles. Technology uses tactile sensations to identify underground objects and researchers think it could be used one day to help disarm terrestrial mines and inspect buried cables. The researchers say it is very difficult to identify articles buried in granular materials.
Identify objects in granular materials with a robot requires fingers sufficiently thin to penetrate the sand, movable enough to undress the material of the search and sufficiently sensitive to feel the detailed form of the buried object. The developed robotic finger trim is equipped with tactile detection that allows him to identify buried objects and is called the finger of Digger.
During the tests, the finger of the hollow can dig through granular materials such as sand or rice and correctly detect the shape of buried objects inside the material. Attempting to identify buried objects inside a granular material, such as sand, gravel and other types of flawless packaged particles is not a new challenge. Researchers used technologies such as radar penetration or ultrasonic vibrations to detect objects buried in the granular form in the past.
These techniques provide a blurred view of submerged objects, making it difficult to differentiate materials, such as saying the difference between the rock and the bone. The team had to conceive the robotic finger to be thin and pointed. In the previous work, the team used a sensor consisting of a clear gel covered with a reflective membrane that can deform when objects were pressed against it called GelSight. Behind the membrane were different colors of LED lights and a camera.
The light shines through the gel and the membrane while the camera collected the pattern of reflection of the membrane. Computer vision algorithms are then extracted from the shape of the contact area where the soft finger touches the object. The device was very good at artificial touch but was very large. The GELSIGHT sensor has been modified in the latest searches changing its shape to a thin cylinder with a beveled tip.
They then removed two thirds from LED lights based on a combination of blue LEDs and colorful fluorescent paint to save complexity and space. The result was the highly functional and capable digital finger sensor. Currently, researchers explore new motions to optimize the ability of the digger to navigate in various media.
