Feel the Force

NYU's low-cost electronic skin called eFlesh gives robots a human-like sense of touch with just a 3D printer and some magnets.

Nick Bild
1 month agoRobotics
This electronic skin is ideal for hobbyist use (📷: V. Pattabiraman et al.)

So you have completed the obligatory first steps in learning about robotics with a line follower and a bump-and-go build — now what comes next? After these initial projects, many hobbyists will be ready to start teaching their robots to interact with the world around them in more meaningful ways. That could mean taking some more baby steps with an ultrasonic sensor or inertial measurement unit, but eventually these robots will need the sense of touch.

Imagine how hard it would be to pick something up or use a tool without this sense. It is no easier for robots lacking this capability. However, adding any type of sensing system that even remotely resembles the function of human skin is traditionally complex and very expensive, leaving the casual hobbyist out in the cold. But a new sensing system developed by engineers at New York University may soon change that.

The team has developed what they call eFlesh (take it easy, they are engineers, not marketers!), which is an inexpensive and simple way to roughly approximate the touch sensing capabilities of human skin. eFlesh can be manufactured for just a few dollars with a 3D printer, some magnets, and a few magnetometers. Because it is 3D-printed, it is also highly customizable, allowing developers to mold it to fit different structures and suit a variety of use cases.

eFlesh mimics touch through magnetic sensing. Unlike traditional tactile sensors, which can be fragile, bulky, or prohibitively expensive, eFlesh is low-cost, simple to build, and adaptable. Its secret lies in a clever design that consists of tiled microstructures printed in a soft material that is embedded with magnets and paired with a magnetometer board. When a force is applied, the material deforms, causing measurable shifts in the magnetic field that can be used to infer the location, direction, and magnitude of contact.

Using the team’s open-source toolchain, any convex 3D shape (in formats like STL or OBJ) can be converted into a printable file that incorporates the necessary structure for sensing. This makes eFlesh a highly modular system that hobbyists, educators, and researchers can tailor for a wide array of robotic hands, grippers, or even prosthetic limbs.

In lab tests, eFlesh demonstrated some impressive accuracy, with contact localization errors under 0.5 mm and force prediction errors of just 0.27 N for normal force and 0.12 N for shear force. The system even supports slip detection, which is crucial for manipulating objects in motion, with a 95% success rate. When compared to vision-only approaches, performance increased by over 40% with the addition of eFlesh.

Whether you are building your first robotic gripper or enhancing an existing platform, the ability to add a robust sense of touch for a few dollars is a big leap forward. eFlesh offers a bridge between high-end research and homegrown experimentation, which finally lets us all get in on the fun.

Nick Bild
R&D, creativity, and building the next big thing you never knew you wanted are my specialties.
Latest articles
Sponsored articles
Related articles
Get our weekly newsletter when you join Hackster.
Latest articles
Read more
Related articles