Inspiration

Healthcare and fitness, especially in these times, need a high tech solution to some pressing issues, especially among the elderly and at-risk populations who may need attention and care but are unable to avail them due to the current conditions

What it does

CryptoFIT is a system where a patient’s health and fitness are monitored actively and tied to their caregivers, physicians, and medical personnel via smart contracts. We have a hardware component which is a smart wearable that detects various health and fitness parameters such as steps taken, average pulse rate, blood oxygen saturation level, body temperature, and galvanic skin response. These readings are stored as transactions on a blockchain that implements a smart contract between a patient and their healthcare provider/caregivers. The contract outlines weekly goals and threshold levels for the patient to achieve and the readings from the wearable device monitors the patient’s progress towards this goal. We incorporate a system to track medication usage by the patient, by using machine learning to identify pills taken by a patient according to their prescription. This forms a secondary contract and the patient is prompted to take a picture of the pill they are about to take just before consuming it. We trained a machine learning model to identify various types of medications, and therefore we also add the functionality to alert patients and caregivers if the wrong medication is accidentally taken, or if the dosage or frequency deviates from the prescription.

How we built it

The CryptoFIT wearable has 2 battery power esp32 microcontroller (M5 StickC and M5 Stack) as its core. These microcontrollers along with multiple sensors take the user’s body vitals such as heart rate, blood oxygen level, body temperature, Galvanic Skin Response, and the number of steps taken. These readings are then stored in an encrypted blockchain as transactions towards a smart contract. The blockchain is running on local nodes, however, we are working on implementing this on a solidity contract on the Ocean Nile Testnet (we have had issues connecting to the Nile testnet)

Challenges we ran into

Combining the MAX10302 sensor with the M5Stack was a cumbersome process MAX10302 libraries use internal Arduino interrupts for blood oxygen level calculation. These interrupts were disabled by the M5Stack internal programming.