Leg Tapping Detection System

Inspiration

You’ll find leg shakes or leg bouncing a commonly referenced symptom of anxiety disorders or a sign that you’re feeling stressed and aren’t aware of it. This has become a common problem for most people and it can disturb us as well as the people working with us. I (Nilutpol) am a leg bouncer. At work, I’m always either sitting with my legs up on the chair in a pretzel-like position, or my feet on the floor but my right leg is furiously shuddering. It can be going for ages without me noticing until I can suddenly feel the shakes all along my desk.

What it does

It detects the shaking or bouncing of our legs using the IMU sensor and based on that it alerts the wearer of their leg shaking using a Red LED and a buzzer alert.

How we built it

We used an ESP32 board for running our code. We attached an MPU6050 IMU for detecting the shaking of the leg. Using the IMU, we collected leg-shaking data for different patterns (up-down shake and left-right shake) and then trained a TinyML model using Edge Impulse and the collected data. The trained TinyML model is deployed to the ESP32 board.

Whenever the device is powered ON, it starts collecting the Acceleration data from the IMU( X, Y, Z) and runs inference on the IMU data, and then classifies it as either shaking or idle. For shaking, it turns on the Red LED and turns ON the Buzzer for the alert.

Then we attached the whole system on the wearer's leg (using tapes for now :‑b).

Challenges we ran into

We faced some trouble in deploying our trained TinyML model onto our ESP32 board due to its less computational power. But we got help from Edge Impulse experts for sorting out the error.