About a decade ago, Ted Adelson’s research group at MTI’s Computer Science and Artificial Intelligence Laboratory (CSAIL) invented a new sensor technology, named GelSight, that used physical contact with an object to provide a significantly detailed 3D map of its surface.

GelSight‘s technology enables digital tactile sensing with the sensitivity and resolution as human touch (hand). The sensor consists of a block of transparent rubber act as skin, called “gel” which on one face of the rubber is coated with metallic paint. When the paint-coated surface is press against an object, it reconcile to the object’s shape.

The metallic paint makes the object’s surface reflective, so its geometry becomes much easier for computer vision algorithms to infer. On the opposite side of the paint-coated rubber, there are three different colored lights and a single camera mounted on the sensor. The camera will capture the image from the pressed “gel” and with the helps of reflected light that being projected from different angles helps the computer to figure out the 3D shape of what that thing is.

For an autonomous robot, gauging object’s softness and hardness is essential to deciding not only where and how hard to grasp them but how they will behave when moved, stacked or laid on different surfaces. Tactile sensing could also aid the robots in differentiating objects that look similar. Previously, in order to determined the hardness of an object, robots have to lay an object on flat surface and gently poking them to see how much pressure they gives.

This is totally different from the way human gauge hardness. Our judgment is based on degree to which the contact area between the object and our fingers change on it. Softer objects tend to flatten more when pressed hence increasing the contact area. The researches adopted the same approach and came out with this idea.

GelSight‘s 3D tactile sensing is a solutions that address a wide variety of surface inspection challenges. The video explains how the use of tactile sensing gives robots added perception and intelligent interaction. This systems could helps in various high-value applications, including;
1. Robotic inspection for quality assurance.
2. Material properties identification.
3. Manufacturing and assembly process efficiency.
4. Industrial inspection to check for scratch, bump or any defect on material surface

In future, GelSight plan to improves the system with temperature and vibration detections to simulate as almost the same as human fingers. With this upgrades, the robot can identify directly more details what type of material it is holding/touching; such as determined whether the thing is metal or wood, is it wet or dry, even can differentiate between unvarnished wood and perfectly smooth surface.

It is not impossible that one day, robot can have fingers exactly the same as us, human, which can sense and moves exactly the same. Besides, GelSight goal is to provide the sensing capability, so that the robot can skillfully interact with the world.

Source: http://www.gelsight.com , MIT News