AKKHIL MORKONDA Projects / Partial Weight Bearing Insole Email me
Back to all projects

Partial Weight Bearing Insole

BMED 2310 design project, Georgia Tech, team of 5, Jan to Jul 2025

What

A wearable that measures how much weight a patient puts on an injured leg and shows it live on their phone, so they stay inside the partial weight bearing limit their physical therapist sets.

Why

After a lower-limb fracture, patients are told to bear only a percentage of their body weight, but even trained physical therapists misjudge it by up to 30%, and training on a bathroom scale does not carry over to walking. The few biofeedback devices cost about $700 to $7,000 or are not sold in the US.

How

  • Placed three 50 kg force-sensitive resistors at the foot's three critical contact points, sewn into a sock liner worn in the patient's own shoe
  • Built the processing unit around an ESP32, with the sensor circuit and a rechargeable 3.7V LiPo with a charging booster, in a 3D-printed PLA housing
  • Wrote the Arduino C++ firmware and calibrated the three sensors with a linear regression against a bathroom scale, across users of different weights
  • Streamed calibrated load over WiFi to Arduino Cloud for a mobile app with a live gauge and progress over time
  • Helped model the looks-like CAD for the next version, a double-layer sock with a miniaturized processing unit
  • ESP32
  • Arduino C++
  • Force-sensitive resistors
  • Calibration
  • Linear regression
  • Arduino Cloud
  • LiPo charging
  • 3D printing
  • CAD
Red 3D-printed PLA housing open, showing the ESP32 processing unit, wiring and battery
Processing unit: ESP32, LiPo and charging in a PLA housing
Foot on a bathroom scale with the phone app reading load during calibration
Calibrating against a bathroom scale
Phone app showing a live load gauge and a load-over-time chart
Live gauge in the Arduino Cloud app
Scatter plot of sensor reading against scale reading at 50% partial weight bearing, with a linear trend line
Sensor vs scale at 50% body weight

Results

  • Within 3% of a reference scale at 25, 50 and 75% body weight (n = 20, paired t-tests)
  • Working prototype from foot sensors to a live phone gauge

Modeling

The CAD for the next version: wiring sealed inside a double-layer sock and a processing unit small enough to sit at the ankle.

CAD render of the double-layer sock with the processing unit at the ankle
Double-layer sock
CAD render of the sock with internal wiring exposed
Internal wiring
CAD render of the sock sole showing sensor placement
Sensor placement
CAD renders of the miniaturized processing unit from several angles
Miniaturized processing unit