Intelligent and intuitive elderly care
Lumo combines AI-powered conversation, health monitoring, and proactive support to enhance the lives of seniors living independently.
Projected elderly population in the US by 2030
of seniors report experiencing isolation
Increased risk of injury when living alone
Designed for independence and safety
Lumo combines cutting-edge technology with intuitive design to provide comprehensive support for seniors living independently.
Adaptive Natural Conversation
Engage in natural, context-aware dialogue with your personal assistant and companion.
Biometric Health Monitoring
Real-time monitoring of vital signs with immediate alerts for any concerning changes.
Proactive Support & Connectivity
Stay connected with family and caregivers through seamless communication and alerts.
Intelligent architecture for seamless care
Lumo's distributed system works together to provide comprehensive monitoring and support.
- 1
Data Capture
Wearable device continuously captures biometric data through advanced sensors.
- 2
Local Processing
Data is immediately processed on-device using TinyML for critical event detection.
- 3
Hub Processing
The Lumo hub receives and further processes data for enhanced health monitoring.
- 4
Cloud Analytics
Long-term data is securely stored and analyzed to identify trends and patterns.
Edge Device
- • Raspberry Pi 5 with BLE module
- • On-device low-power DSP
- • Lightweight wake word engine
- • OpenAI Realtime API integration
- • Preprocessed sensor data retrieval
Wearable Device
- • Arduino Nano 33 Rev2
- • MAX30102 & MAX30205 sensors
- • Interrupt-driven routines
- • Local anomaly detection
- • Bluetooth Low Energy (BLE)
Data Pipeline
- • I2C bus communication at 50Hz
- • Circular buffer for data bursts
- • Real-time data streaming
- • TinyML for time-critical detection
- • SupaBase cloud storage
- • Historical data aggregation
Technical challenges overcome
Building Lumo required solving complex technical problems to ensure reliable, real-time performance.
Maintaining Stable Sampling Rate
While integrating the two biosensors on the wearable device, I observed that the intended 50Hz sampling rate was inconsistent, with sporadic delays that led to dropped sensor readings.
Solution:
- Revised approach from polling-based loop to hardware interrupts
- Configured data-ready pins on each sensor
- Implemented FreeRTOS task scheduling
- Optimized timer management for consistent sampling
Rapid Anomaly Detection
Processing incoming sensor data in real time on the Raspberry Pi-powered hub device was challenging: the data needed to be quickly filtered, normalized, and analyzed in order to trigger immediate and timely alerts.
Solution:
- Implemented multi-threaded processing within the hub device
- Separated local data processing, voice interaction, and BLE data reception tasks
- Optimized data pipeline with asynchronous I/O using libevent
- Significantly reduced latency between anomaly detection and alert generation
Project timeline and milestones
Follow Lumo's development journey from concept to launch.
Q4 2024
Research & Planning
Initial concept development, market research, and technology evaluation.
Q1 2025
Architecture Design & Hardware Prototyping
System architecture development and initial hardware prototype creation.
Q2 2025
Software/Firmware Development
Development of core software components, firmware, and integration testing.
Q3 2025
Hardware Reworking & Optimization
Final hardware refinements, debugging, and performance optimization.
Current Status
Lumo is currently in the development phase, with active work on firmware and software integration on the prototype. We're focused on optimizing the data pipeline and improving the reliability of health monitoring algorithms.
Ready to enhance independent living?
Join the Lumo family and experience the future of elderly care today.
We'll send you information about Lumo and how to get started. Privacy Policy