All Case Studies
Healthcare & Medical Devices
14 months
PulseGuard Health Technologies

Medical Grade Wearable for Continuous Cardiac Monitoring

Medical Grade Wearable for Continuous Cardiac Monitoring
Overview

Project Overview

EmbedCrest, an embedded systems company in India specializing in embedded & engineering services, partnered with PulseGuard Health Technologies to engineer a medical grade continuous cardiac monitoring wearable for post-surgical patient care. The device was built on the Nordic nRF5340 dual core SoC running Zephyr RTOS, integrating a Texas Instruments ADS1293 24-bit analog front end for clinical quality three lead ECG acquisition and a Maxim MAX30102 optical sensor for SpO2 monitoring. Our edge AI development services team built a custom 1D convolutional neural network quantized to 8-bit integers running on the ARM Cortex-M33 application core, performing real time classification of atrial fibrillation, premature ventricular contractions, and bradycardia with 98.3 percent sensitivity. BLE 5.3 with coded PHY ensured reliable data transmission to a companion mobile application, while AES-256 encryption protected all patient health data in transit and at rest. The entire power optimized firmware development followed IEC 62304 Class B software lifecycle processes with complete requirements traceability, supporting PulseGuard preparation for FDA 510(k) submission. Delivered from our engineering center in Mumbai, India, the final device weighed 28 grams and achieved 8.2 days of continuous monitoring on a single charge.

The Challenge

Problem Statement

PulseGuard Health Technologies required a clinical grade wearable device for continuous ECG and SpO2 monitoring in post-surgical cardiac patients. The device needed to achieve medical accuracy comparable to Holter monitors, provide real time arrhythmia detection with FDA 510(k) submission readiness, and deliver 7 day continuous battery life in a form factor weighing under 35 grams. They were looking for embedded systems companies in India with proven embedded & engineering services in the medical device domain.

Our Solution

Engineering Approach

EmbedCrest designed the wearable around a Nordic nRF5340 dual core SoC, with the network core handling BLE 5.3 connectivity and the application core running Zephyr RTOS for deterministic real time ECG signal acquisition. A Texas Instruments ADS1293 analog front end provided clinical grade 24-bit ECG capture across three leads, while a Maxim MAX30102 optical sensor handled SpO2 and heart rate measurements. An on device arrhythmia detection algorithm based on a 1D convolutional neural network, quantized to 8-bit integers and running on the ARM Cortex-M33, classified atrial fibrillation, premature ventricular contractions, and bradycardia events with clinical grade sensitivity. BLE 5.3 with coded PHY enabled reliable data streaming to a companion mobile application, which relayed critical alerts to the clinical monitoring platform over HTTPS. The firmware implemented IEC 62304 compliant software lifecycle processes, with full traceability from requirements to test cases. Our embedded developers in India delivered the complete power optimized firmware development and edge AI inference pipeline.

Results

Measurable Outcomes

98.3%

Arrhythmia detection sensitivity

8.2 days

Continuous battery life

28 grams

Device weight

Achieved

FDA 510(k) submission readiness

Tech Stack

Technologies Used

Nordic nRF5340Zephyr RTOSBLE 5.3TI ADS1293MAX30102TensorFlow Lite MicroARM Cortex-M33IEC 62304AES-256 encryptionFHIR HL7 (cloud)Coded PHY
EmbedCrest understood the regulatory rigor required for medical devices from day one. Their firmware team delivered clinical grade ECG acquisition quality that matched our gold standard Holter reference, and the on device arrhythmia detection exceeded our accuracy targets. They were instrumental in preparing our 510(k) technical documentation.

Dr. Vikram Rao

Chief Medical Officer, PulseGuard Health Technologies

Note: Client details have been anonymized to protect confidentiality. Project outcomes and technical specifications are representative of actual engagements.

FAQs

Frequently Asked Questions

How does the device achieve medical grade ECG accuracy?

The Texas Instruments ADS1293 provides 24-bit resolution analog to digital conversion with integrated right leg drive and lead off detection. Combined with a custom analog signal conditioning circuit featuring active shielding and a 0.05Hz to 150Hz bandpass filter, the system achieves noise levels below 10 microvolts RMS, meeting IEC 60601-2-47 requirements for ambulatory ECG devices.

What regulatory standards does the firmware comply with?

The firmware was developed following IEC 62304 Class B software lifecycle processes, with full traceability from system requirements through software architecture, unit tests, and integration tests. Risk management followed ISO 14971, and the ECG measurement subsystem was validated against IEC 60601-2-47. All documentation was structured to support FDA 510(k) submission.

How is patient data protected on the device?

Patient data is encrypted at rest using AES-256 with keys stored in the nRF5340 hardware key management unit. BLE communication uses secure connections with LESC pairing and additional application layer encryption. The device stores no personally identifiable information, using only anonymized session tokens that are linked to patient records exclusively on the HIPAA compliant cloud platform.

Ready to Build Something Remarkable?

Let's discuss how our embedded systems expertise can deliver measurable results for your project.