# The robotic hand (2016–2017)

> The last-year-of-high-school project in Italy: a robotic hand that mimicked a human hand through a glove embedded with flex sensors, built close to the bare metal to bring together everything Massimo had learned about electronics.

- Canonical URL: [https://massimo.context-use.com/p/about/projects/2016-2017_robotic-hand/story](https://massimo.context-use.com/p/about/projects/2016-2017_robotic-hand/story)
- Site index: [https://massimo.context-use.com/llms.txt](https://massimo.context-use.com/llms.txt)
- Last edited: 2026-07-26T11:09:08.503Z

After coming back from my [exchange year in the US](https://massimo.context-use.com/p/about/chapters/2015-2016_novato/story), I was charged with ambition. My last year of high school was coming up (Italy has five years of it, unfortunately), and I wanted to build another project — one that would bring together all the electronics knowledge I had gathered [throughout high school](https://massimo.context-use.com/p/about/chapters/1998-2017_como/story).

This time, instead of relying on a Raspberry Pi, I wanted to get closer to the bare metal and build something great. That's how the robotic hand came to be.

![The robotic hand system: the 3D-printed hand with its servos, the custom control board, and the sensor glove](https://assets.massimo.context-use.com/a/about/projects/2016-2017_robotic-hand/hero){size=full align=center}

Together with a classmate of mine (who would eventually go to [Politecnico](https://massimo.context-use.com/p/about/chapters/2017-2020_politecnico/story) with me), we decided to build a robotic hand that could mimic the movements of a human hand thanks to a glove embedded with flex sensors. These are simple devices whose electric resistance changes proportionally to the degree to which they are bent. Feeding a current into the flex sensor caused a voltage to appear across its two ends; bending the flex sensor changed the resistance — and with it, according to Ohm's law, the voltage. The resulting signal was then processed and used to drive five servo motors, one for each finger.

![The glove, with a flex sensor running along each finger](https://assets.massimo.context-use.com/a/about/projects/2016-2017_robotic-hand/glove-flex-sensors){size=large align=center}

The voltage measured from each sensor was then fed into an operational amplifier to adjust the signal, before going into an [ATmega328](https://en.wikipedia.org/wiki/ATmega328) microcontroller. The latter had a bunch of A/D converters that allowed us to turn the analog signal into a digital one, which the processor could understand and use to drive the motors attached to each finger of the robotic hand.

We designed the PCB ourselves and etched the copper using some acids that are better to stay away from. We messed up the circuit a few times, but managed to get away with a functioning board.

![The home-etched control board: the microcontroller in the middle, one operational amplifier per finger on the right](https://assets.massimo.context-use.com/a/about/projects/2016-2017_robotic-hand/control-board){size=large align=center}

Thankfully, by 2016 3D printers were booming, and we found an open-source model of a prosthetic hand which we used as our "robotic hand". Our mechanical skills were not that sophisticated, so we focused only on the electronics side. Our mechanical "innovations" were using a printable elastomer material for the joints (so the fingers could bend) and gluing simple elastics on top of the fingers (so they could get back to their natural position). This allowed us to use simple [SG90 servo motors](http://www.ee.ic.ac.uk/pcheung/teaching/DE1_EE/stores/sg90_datasheet.pdf) to bend the fingers using fishing lines — the elastics would take care of bringing each finger back as soon as the servo moved back to its initial position.

![The 3D-printed hand from the back, with the five SG90 servos and their fishing lines](https://assets.massimo.context-use.com/a/about/projects/2016-2017_robotic-hand/hand-with-servos){layout=half}
![A printed finger, with the elastomer joints in blue](https://assets.massimo.context-use.com/a/about/projects/2016-2017_robotic-hand/printed-finger-joints){layout=half}

The robotic hand was basically useless: it could barely hold any weight, and its movements were really limited. But the fact that we had built it from the ground up — and that it was able to replicate our hand movements in real time — made it fucking amazing.

![Demo: the robotic hand replicating hand movements in real time](https://assets.massimo.context-use.com/a/about/projects/2016-2017_robotic-hand/demo-video-1){layout=third size=small}
![Demo: driving the fingers through the sensor glove](https://assets.massimo.context-use.com/a/about/projects/2016-2017_robotic-hand/demo-video-2){layout=third size=small}
![Demo: the hand in action](https://assets.massimo.context-use.com/a/about/projects/2016-2017_robotic-hand/demo-video-3){layout=third size=small}

It's fun to see that, even though ours was little more than a toy project, some of today's best robotics companies have created robotic hands that do not look that different — see [mimic robotics](https://www.mimicrobotics.com/blog/solving-dexterity-a-full-stack-approach) or [1X's NEO hands](https://www.1x.tech/discover/neos-hands). (Okay, okay — they are quite different. But come on, we were two high school students with little idea of what we were doing.)

After this project, I felt like I was [ready to go and build rockets](https://massimo.context-use.com/p/about/projects/2017_building-rockets/story)! Unfortunately though, I went to [Politecnico di Milano](https://massimo.context-use.com/p/about/chapters/2017-2020_politecnico/story), and my side projects came to a halt for almost three years :(