Pressure Injury Prevention System 2.0

PCB Design

I designed two complementary PCBs that form the hardware backbone of PIPS 2.0:

Flexible PCB for Sensor Integration

This flexible PCB is sewn directly onto the textile material containing the pressure and moisture sensors. The flexible design allows the PCB to conform to the sensor matrix geometry without introducing mechanical stress. It features optimized routing to minimize cross-talk between analog sensor signals, ensuring clean data acquisition from the piezoresistive array.

Main Acquisition Board

A four-layer rigid PCB serves as the data acquisition hub. It houses an ESP32 microcontroller that gathers analog data from one section of the sensor mat through flexible PCBs. The board includes signal conditioning circuitry for filtering and amplification, ensuring robust analog-to-digital conversion. Communication with the central Raspberry Pi occurs over I2C, allowing multiple acquisition boards to operate simultaneously on the same bus for scalability.

Key design considerations: impedance matching, thermal management, and noise isolation between analog and digital domains.

Firmware for ESP32 Microcontroller

I developed embedded firmware for the ESP32 that handles real-time sensor data acquisition with dual transmission modes:

Data Acquisition Pipeline

The firmware implements a high-frequency ADC sampling routine that continuously reads the pressure and moisture sensor values at configurable intervals. Raw analog data undergoes preliminary buffering and timestamping for accurate temporal correlation in downstream processing.

Dual Transmission Modes

I2C Mode (Primary): Sends processed sensor data to the Raspberry Pi over I2C protocol. The ESP32 acts as an I2C slave device, making it queryable by the central controller. This mode enables tight synchronization and centralized data management.

BLE Mode (Secondary): Provides an alternative wireless transmission path using Bluetooth Low Energy. This enables direct communication with mobile devices or backup connectivity if I2C fails, adding resilience to the system.

The firmware also implements error handling, watchdog timers, and power management features to ensure long-term reliability in a clinical environment.

Backend Software - Python Server

I built a Python-based backend server that aggregates data from multiple ESP32 microcontrollers and orchestrates the complete data pipeline:

Multi-Device Data Aggregation

The server communicates with multiple ESP32 acquisition boards simultaneously over I2C, collecting sensor data at regular intervals. It manages device discovery, connection health monitoring, and automatic reconnection logic to handle network disruptions gracefully.

Compensation Algorithms

Raw sensor data contains artifacts from thermal drift, sensor aging, and environmental factors. I implemented sophisticated compensation algorithms that apply calibration corrections and normalization to transform raw readings into clinically meaningful pressure and moisture values. These algorithms are continuously refined based on ground-truth measurements.

Data Logging and Firebase Integration

The server maintains local data logs for offline resilience and redundancy. Processed data is pushed to Firebase in real-time, enabling cloud storage, remote monitoring, and integration with clinical dashboards. The architecture supports historical queries and statistical analysis for long-term patient monitoring trends.

Additional features include API endpoints for configuration, data export, and system diagnostics to support clinical integration and research workflows.