Skip to main content

GROOT FORCE - FRD: Sensor & Safety Systems

Document Version: 1.0
Date: November 2025
Status: Active Development
Classification: Internal - Engineering

Document Control​

VersionDateAuthorChanges
1.0Nov 2025Engineering TeamInitial sensor & safety FRD

Approval:

  • Hardware Lead: _________________ Date: _______
  • Safety Engineer: _________________ Date: _______
  • Software Architect: _________________ Date: _______
  • Product Manager: _________________ Date: _______

1. Executive Summary​

1.1 Purpose​

This Functional Requirements Document (FRD) defines the complete sensor suite and safety intelligence system for GROOT FORCE smart glasses. These systems enable:

  • Walking assistance and obstacle detection
  • Fall detection and prevention
  • Environmental awareness
  • Health monitoring
  • Gesture recognition
  • Spatial mapping and AR accuracy

The sensor and safety systems are mission-critical - they directly impact user safety, particularly for NDIS assistive variants and enterprise industrial use.

1.2 Scope​

In Scope:

  • Motion sensors (IMU - accelerometer, gyroscope, magnetometer)
  • Depth sensors (ToF, LiDAR)
  • Environmental sensors (BME688 - temperature, humidity, VOC/gas)
  • Health sensors (MAX30102 - heart rate, SpO₂; MLX90614 - temperature)
  • Sensor fusion algorithms
  • Walking assistance system
  • Fall detection system
  • Obstacle detection and alerts
  • Health monitoring and alerts
  • Safety intelligence engine
  • Calibration procedures
  • Sensor HAL (Hardware Abstraction Layer)

Out of Scope:

  • Camera systems (covered in separate FRD)
  • Audio systems (covered in separate FRD)
  • Display systems (covered in Hardware Requirements)
  • AI processing (covered in FRD: Core AI System)

2. System Architecture Overview​

2.1 Sensor System Hierarchy​

┌─────────────────────────────────────────────────────────┐
│                    KLYRA OS / App Layer                 │
│         (Walking Assist, Fall Detection, Health)        │
└────────────────────┬────────────────────────────────────┘
                     │
┌────────────────────▼────────────────────────────────────┐
│              Sensor Fusion Engine                       │
│    (Kalman Filters, Quaternions, Data Correlation)     │
└────────────────────┬────────────────────────────────────┘
                     │
┌────────────────────▼────────────────────────────────────┐
│              Sensor HAL (Hardware Abstraction)          │
│         (Standardized interfaces, calibration)          │
└────┬───────┬───────┬───────┬───────┬───────────────────┘
     │       │       │       │       │
┌────▼──┐┌──▼───┐┌──▼───┐┌──▼───┐┌──▼────┐
│  IMU  ││ ToF  ││LiDAR ││ BME  ││Health │
│9-axis ││Depth ││Long  ││Enviro││HR/SpO2│
└───────┘└──────┘└──────┘└──────┘└───────┘

2.2 Data Flow​

  1. Raw Sensor Data → 2. HAL Processing → 3. Fusion Engine → 4. Application Logic → 5. User Alerts/Actions

3. Motion Sensors (IMU)​

3.1 Overview​

Purpose: Track head orientation, movement, gestures, detect falls, enable AR stability

Hardware: ICM-20948 9-axis IMU (or equivalent)

  • 3-axis accelerometer
  • 3-axis gyroscope
  • 3-axis magnetometer

3.2 Functional Requirements​

FRD-SENS-IMU-001: Continuous Motion Tracking​

Description: The IMU shall continuously track device orientation and motion at minimum 200 Hz sampling rate.

Inputs:

  • Raw accelerometer data (±16g range)
  • Raw gyroscope data (±2000 dps range)
  • Raw magnetometer data (±4900 µT range)

Outputs:

  • 6DoF pose (position + orientation)
  • Quaternion representation
  • Linear acceleration
  • Angular velocity

Performance:

  • Sampling rate: 200-1000 Hz (configurable)
  • Latency: < 5ms from sensor to HAL
  • Drift correction: < 0.5° per minute

Validation:

  • Unit test: IMU data streaming at target rate
  • Integration test: Pose accuracy vs ground truth
  • Field test: Long-term drift measurement

Priority: P0 (Critical)

FRD-SENS-IMU-002: Gesture Recognition​

Description: The IMU shall detect specific head gestures for user input.

Supported Gestures:

  • Nod (yes): 2 quick forward tilts
  • Shake (no): 2 quick left-right turns
  • Tilt up: Look up > 30°
  • Tilt down: Look down > 30°

Performance:

  • Recognition accuracy: > 95%
  • False positive rate: < 2%
  • Latency: < 200ms from gesture start to detection

Validation:

  • User testing with 50+ participants
  • Edge case testing (walking, running, driving)
  • Confusion matrix analysis

Priority: P1 (High)

FRD-SENS-IMU-003: Fall Detection​

Description: The system shall detect when the user is falling and trigger alerts.

Detection Algorithm:

  1. Monitor acceleration magnitude
  2. Detect freefall: |a| < 0.5g for > 300ms
  3. Detect impact: |a| > 2.5g
  4. Check orientation change: > 45° tilt
  5. Check stillness after impact: < 0.2g for > 3s

Outputs:

  • Fall event detected (boolean)
  • Confidence score (0-100%)
  • Pre-fall/post-fall timestamps
  • Fall trajectory data

Actions:

  • Immediate haptic alert
  • Audio alert: "Fall detected. Are you okay?"
  • If no response in 30s: Send emergency alert to contacts
  • Log fall event to health history

Performance:

  • Detection accuracy: > 90%
  • False positive rate: < 5%
  • Time to alert: < 2s from fall start

Validation:

  • Controlled fall simulations with crash test dummies
  • User acceptance testing with NDIS participants
  • Edge case testing: sitting down quickly, lying down intentionally

Priority: P0 (Critical - NDIS safety feature)

FRD-SENS-IMU-004: Activity Classification​

Description: The IMU shall classify user activity state.

Activity States:

  • Stationary (standing/sitting)
  • Walking (1-3 Hz gait)
  • Running (3-6 Hz gait)
  • Cycling (5-30 Hz vibration)
  • Driving ( > 10 km/h velocity)

Outputs:

  • Current activity state
  • Confidence level
  • State duration
  • Transition timestamps

Use Cases:

  • Enable driving mode (disable distracting features)
  • Adjust walking assist sensitivity
  • Adapt AI response length (shorter during movement)

Performance:

  • Classification accuracy: > 85%
  • State transition latency: < 3s

Validation:

  • Dataset collection across all activity types
  • Real-world testing in various environments

Priority: P1 (High)

FRD-SENS-IMU-005: Calibration​

Description: The IMU shall support field calibration to compensate for manufacturing variance and magnetic interference.

Calibration Procedures:

  1. Accelerometer bias calibration (device level on surface)
  2. Gyroscope bias calibration (device stationary)
  3. Magnetometer calibration (figure-8 motion)

Automatic Calibration:

  • Detect when device is stationary for > 5s
  • Auto-calibrate accelerometer and gyroscope bias
  • Store calibration offsets in persistent storage

Manual Calibration:

  • User-initiated via settings menu
  • Guided on-screen instructions
  • Pass/fail indication

Validation:

  • Verify calibration improves accuracy by > 20%
  • Test across temperature range: -10°C to +50°C

Priority: P1 (High)

4. Depth Sensors​

4.1 Time-of-Flight (ToF) Sensor​

FRD-SENS-TOF-001: Short-Range Obstacle Detection​

Description: The ToF sensor shall provide short-range depth mapping for obstacle detection and hand tracking.

Hardware: VL53L5CX (or equivalent)

  • 8×8 multi-zone ranging
  • Range: 0.2m - 4m
  • Field of view: ~45°

Inputs:

  • IR laser pulses
  • Ambient light conditions

Outputs:

  • 8×8 depth map (64 zones)
  • Per-zone distance (mm)
  • Per-zone confidence
  • Ambient light level

Performance:

  • Update rate: 15-60 Hz
  • Accuracy: ±3% at 1m
  • Min distance: 20cm
  • Max distance: 4m

Use Cases:

  • Detect obstacles directly in front (walls, doors, furniture)
  • Hand gesture tracking (swipe, wave)
  • Ground mapping for walking assist (short range)

Validation:

  • Accuracy test vs laser rangefinder
  • Edge case testing: glass, mirrors, dark surfaces
  • Integration test with walking assist

Priority: P0 (Critical for walking assist)

FRD-SENS-TOF-002: Hand Gesture Tracking​

Description: The ToF sensor shall enable hands-free gesture control.

Supported Gestures:

  • Swipe left/right: Hand moves laterally across FOV
  • Wave: Hand moves toward/away from sensor repeatedly
  • Air tap: Quick forward motion and retract
  • Hold: Hand stationary in specific zone for 2s

Performance:

  • Gesture detection rate: > 90%
  • Latency: < 150ms
  • Operating range: 20cm - 60cm from glasses

Validation:

  • User testing with diverse hand sizes
  • Lighting condition testing (indoor, outdoor, darkness)

Priority: P1 (High)

4.2 LiDAR Sensor​

FRD-SENS-LIDAR-001: Long-Range Spatial Mapping​

Description: The LiDAR sensor shall provide long-range distance measurement for navigation and safety.

Hardware: TFmini-S (or equivalent)

  • Range: 0.3m - 12m
  • Beam divergence: 0.03°
  • Eye-safe: Class 1

Inputs:

  • Laser pulse timing
  • Target reflectance

Outputs:

  • Distance measurement (cm)
  • Signal strength
  • Measurement confidence

Performance:

  • Update rate: 100 Hz
  • Accuracy: ±6cm at 3m, ±2% at > 3m
  • Resolution: 1cm

Use Cases:

  • Detect distant obstacles (vehicles, poles, walls)
  • Measure room dimensions
  • Navigation path clearance
  • Industrial measurement mode

Validation:

  • Accuracy test vs surveying equipment
  • Safety test for eye exposure
  • Environmental testing (rain, fog, dust)

Priority: P0 (Critical for walking assist)

FRD-SENS-LIDAR-002: Ground Hazard Detection​

Description: The LiDAR shall detect ground-level hazards (steps, curbs, holes).

Detection Algorithm:

  1. Aim LiDAR at ground 2-4m ahead
  2. Build ground profile over time
  3. Detect sudden elevation changes > 5cm
  4. Classify hazard type: step up, step down, hole, slope

Outputs:

  • Hazard detected (boolean)
  • Hazard type (enum)
  • Distance to hazard (m)
  • Severity score (0-100)

Actions:

  • Haptic warning (vibration pattern based on severity)
  • Audio alert: "Step ahead" or "Curb ahead"
  • HUD visual indicator (if enabled)

Performance:

  • Detection range: 2-8m
  • Detection accuracy: > 85%
  • False positive rate: < 10%
  • Warning time: > 3s before hazard

Validation:

  • Real-world testing on various surfaces
  • Edge case testing: ramps, shadows, textures
  • NDIS user acceptance testing

Priority: P0 (Critical - core safety feature)

5. Environmental Sensors​

5.1 BME688 Environmental Sensor​

FRD-SENS-ENV-001: Environmental Monitoring​

Description: The BME688 sensor shall monitor environmental conditions for user comfort and safety.

Measurements:

  • Temperature: -40°C to +85°C (±1°C accuracy)
  • Humidity: 0-100% RH (±3% accuracy)
  • Pressure: 300-1100 hPa (±1 hPa accuracy)
  • Gas resistance (VOC detection): 0-500 kΩ

Outputs:

  • Current temperature (°C)
  • Current humidity (%)
  • Current pressure (hPa)
  • Air quality index (0-500)
  • VOC detected (boolean)

Performance:

  • Update rate: 1 Hz (continuous monitoring)
  • Response time: < 1s (temperature/humidity), < 5s (VOC)

Use Cases:

  • Heat stress warning ( > 35°C + > 70% humidity)
  • Cold exposure warning ( < 0°C)
  • Poor air quality alert (VOC detected, AQI > 150)
  • Altitude estimation (pressure-based)

Validation:

  • Calibration vs certified environmental chamber
  • Real-world testing in various climates
  • VOC detection test with known substances (isopropyl alcohol, acetone)

Priority: P2 (Medium - nice to have)

FRD-SENS-ENV-002: Hazardous Gas Detection​

Description: The system shall alert users to potentially hazardous gases.

Detectable Gases (via VOC sensor pattern matching):

  • Alcohols (ethanol, isopropanol)
  • Acetone
  • Solvents
  • Smoke particles (indirect - via air quality change)

Alert Thresholds:

  • VOC resistance change > 30%: Warning
  • VOC resistance change > 50%: Danger alert
  • AQI > 200: Immediate evacuation recommendation

Actions:

  • Haptic + audio alert
  • HUD warning icon
  • Log exposure event

Limitations:

  • Not a certified gas detector
  • Cannot detect CO, CO₂, natural gas directly
  • User advised to use proper PPE in industrial settings

Priority: P2 (Medium - safety enhancement, not primary safety feature)

6. Health Sensors​

6.1 MAX30102 Heart Rate & SpO₂ Sensor​

FRD-SENS-HEALTH-001: Heart Rate Monitoring​

Description: The MAX30102 sensor shall measure user heart rate via photoplethysmography (PPG).

Hardware:

  • Dual LED (red + infrared)
  • Photodetector
  • Location: Nose bridge or temple (skin contact point)

Outputs:

  • Heart rate (BPM): 30-220 BPM range
  • Heart rate variability (HRV): SDNN, RMSSD
  • Signal quality indicator (0-100%)
  • Confidence level

Performance:

  • Update rate: 1 Hz (1-second moving average)
  • Accuracy: ±3 BPM vs. medical-grade ECG
  • Measurement time: 5-10s for initial reading

Use Cases:

  • Fitness tracking (GF-NF)
  • Stress detection (elevated HR + context)
  • Fatigue monitoring (reduced HRV)
  • Medical monitoring (GF-CL NDIS, with disclaimers)

Validation:

  • Clinical testing vs. Polar H10 chest strap
  • Motion artifact testing (walking, running)
  • Skin tone testing (diverse participants)

Priority: P1 (High for fitness/health variants)

FRD-SENS-HEALTH-002: Blood Oxygen (SpO₂) Monitoring​

Description: The MAX30102 sensor shall measure blood oxygen saturation.

Outputs:

  • SpO₂ percentage: 70-100% range
  • Perfusion index (signal strength)
  • Confidence level

Performance:

  • Update rate: 1 Hz
  • Accuracy: ±2% vs. medical oximeter at 70-100% SpO₂
  • Measurement time: 10-15s

Use Cases:

  • High-altitude monitoring (GF-TR travel, GF-NF fitness)
  • Sleep apnea detection (overnight trends)
  • COVID-19 monitoring
  • Medical monitoring (GF-CL NDIS, with disclaimers)

Medical Disclaimer:

  • Not a certified medical device
  • Results for wellness tracking only
  • User advised to consult doctor for medical decisions

Validation:

  • Clinical testing vs. Masimo medical oximeter
  • Altitude chamber testing
  • Motion artifact testing

Priority: P1 (High for health variants)

FRD-SENS-HEALTH-003: Fatigue Detection​

Description: The system shall detect user fatigue and recommend rest.

Detection Algorithm:

  1. Monitor HRV trends (decreasing HRV = fatigue)
  2. Monitor movement patterns (reduced activity = fatigue)
  3. Monitor voice tone via microphones (flat affect = fatigue)
  4. Combine signals with time-of-day context

Fatigue Levels:

  • Normal: HRV within personal baseline
  • Mild fatigue: HRV 10-20% below baseline
  • Moderate fatigue: HRV 20-40% below baseline
  • Severe fatigue: HRV > 40% below baseline

Actions:

  • Mild: Gentle notification "Consider taking a break"
  • Moderate: Strong recommendation "You seem tired, rest soon"
  • Severe: Insistent alert "Fatigue detected, please rest now"

Use Cases:

  • FIFO workers (GF-TX) - prevent accidents
  • Drivers - suggest pull over
  • NDIS users (GF-CL) - prevent exhaustion
  • Athletes (GF-NF) - optimize training recovery

Validation:

  • Correlation study vs. self-reported fatigue
  • Sleep deprivation testing
  • Post-exercise testing

Priority: P1 (High - safety feature)

6.2 MLX90614 Infrared Temperature Sensor​

FRD-SENS-HEALTH-004: Body Temperature Monitoring​

Description: The MLX90614 sensor shall measure forehead skin temperature.

Hardware:

  • Non-contact IR thermometer
  • Location: Bridge area (faces forehead)
  • Range: -40°C to +125°C (object temp)

Outputs:

  • Skin temperature (°C)
  • Estimated core temperature (compensated)
  • Fever alert ( > 37.5°C core estimate)

Performance:

  • Accuracy: ±0.5°C
  • Update rate: 1 Hz
  • Measurement time: < 1s

Use Cases:

  • Fever detection (GF-CL NDIS, health tracking)
  • Heat stress monitoring (GF-TX industrial)
  • Wellness screening

Medical Disclaimer:

  • Not a diagnostic device
  • Skin temperature ≠ actual core body temperature
  • User advised to use medical thermometer if fever suspected

Validation:

  • Calibration vs. clinical forehead thermometer
  • Ambient temperature compensation testing

Priority: P2 (Medium - nice to have)

7. Sensor Fusion Engine​

7.1 Overview​

The sensor fusion engine combines data from multiple sensors to create a unified, accurate understanding of the user and environment.

FRD-SENS-FUSION-001: 6DoF Pose Estimation​

Description: The fusion engine shall estimate device pose using IMU + visual odometry (when camera available).

Inputs:

  • IMU: acceleration, angular velocity
  • Camera: optical flow (optional)
  • Magnetometer: heading reference

Algorithm: Extended Kalman Filter (EKF) or Madgwick filter

Outputs:

  • Position (x, y, z)
  • Orientation (quaternion or Euler angles)
  • Velocity
  • Pose confidence

Performance:

  • Update rate: 100 Hz
  • Drift: < 0.5° per minute (with mag), < 5° per minute (IMU-only)

Priority: P0 (Critical for AR and walking assist)

FRD-SENS-FUSION-002: Obstacle Map Generation​

Description: The fusion engine shall build a real-time obstacle map using ToF + LiDAR + IMU.

Inputs:

  • ToF: 8×8 depth map (short range)
  • LiDAR: Single-point distance (long range)
  • IMU: Device orientation
  • Camera: Visual features (optional)

Processing:

  1. Transform sensor data to world coordinates using IMU pose
  2. Build 3D point cloud
  3. Segment obstacles vs. ground
  4. Classify obstacle types: wall, furniture, person, vehicle
  5. Maintain obstacle map (5m × 5m around user)

Outputs:

  • Obstacle map (2D grid: 50cm resolution)
  • Obstacle list (type, distance, direction)
  • Clear path indicator (boolean + direction)

Performance:

  • Map update rate: 10 Hz
  • Map accuracy: ±15cm
  • Obstacle classification accuracy: > 70%

Use Cases:

  • Walking assist navigation
  • AR placement of virtual objects
  • Industrial safety (GF-TX)

Priority: P0 (Critical for walking assist)

FRD-SENS-FUSION-003: Context-Aware Sensor Optimization​

Description: The fusion engine shall dynamically adjust sensor sampling rates and processing based on context.

Optimization Rules:

ContextIMU RateToF RateLiDAR RateHealth RateReason
Stationary200 Hz5 Hz10 Hz1 HzMinimal motion, save power
Walking500 Hz30 Hz100 Hz0.2 HzWalking assist priority
Running500 Hz15 Hz50 Hz0.5 HzReduce processing, focus on motion
Cycling200 Hz5 Hz20 Hz0.5 HzLess walking assist, some depth
Driving100 Hz1 Hz1 Hz0.1 HzMinimal sensing, focus on safety
VR/AR Mode1000 Hz60 Hz100 Hz0.1 HzMaximum AR accuracy

Benefits:

  • Extended battery life (up to 30% savings)
  • Reduced thermal load
  • Optimized processing resources

Priority: P1 (High - power efficiency)

8. Walking Assistance System​

8.1 System Overview​

The walking assistance system is the flagship safety feature for NDIS and assistive variants. It helps users navigate safely by detecting obstacles, ground hazards, and providing guidance.

FRD-WALK-001: Real-Time Obstacle Detection​

Description: The system shall detect obstacles in the user's walking path.

Detection Algorithm:

  1. Combine ToF short-range + LiDAR long-range
  2. Build obstacle map (0.2m - 8m ahead)
  3. Analyze walking direction (from IMU)
  4. Predict collision path
  5. Classify obstacle severity (minor, moderate, critical)

Alert Zones:

  • Safe: > 3m to obstacle, no alert
  • Caution: 2-3m to obstacle, gentle haptic pulse
  • Warning: 1-2m to obstacle, strong haptic + audio beep
  • Danger: < 1m to obstacle, continuous haptic + audio "STOP"

Obstacle Types:

  • Walls, doors, furniture (static)
  • People, animals, vehicles (moving)
  • Overhanging objects (tree branches, signs)

Performance:

  • Detection range: 0.2m - 8m
  • Detection accuracy: > 90%
  • False positive rate: < 8%
  • Alert latency: < 300ms

Validation:

  • Obstacle course testing with NDIS participants
  • Various lighting conditions (bright sun, darkness, rain)
  • Edge cases: glass doors, black surfaces, mirrors

Priority: P0 (Critical - core NDIS feature)

FRD-WALK-002: Ground Hazard Detection​

Description: The system shall detect ground-level hazards (steps, curbs, holes, slopes).

Detection Algorithm:

  1. LiDAR scans ground 2-4m ahead
  2. Build ground height profile
  3. Detect elevation changes > 5cm
  4. Classify hazard type

Hazard Types:

  • Step up: Ground rises > 10cm
  • Step down: Ground drops > 10cm (more dangerous)
  • Curb: Sharp ground transition 10-20cm
  • Hole: Ground depression > 20cm
  • Slope: Gradual elevation change > 15°

Alert System:

  • Step up: Haptic + "Step up ahead"
  • Step down: Strong haptic + "Step down, careful!"
  • Curb: Haptic + "Curb ahead"
  • Hole: Strong haptic + "Hazard ahead, STOP"
  • Slope: Gentle haptic (informational)

HUD Overlay (optional):

  • Draw ground profile line on lens
  • Color-code hazards: yellow (caution), red (danger)

Performance:

  • Detection range: 2-8m
  • Detection accuracy: > 85%
  • Warning time: > 3s before hazard

Validation:

  • Real-world testing on stairs, curbs, potholes
  • Comparison to white cane detection
  • NDIS user feedback sessions

Priority: P0 (Critical - core NDIS feature)

FRD-WALK-003: Navigation Guidance​

Description: The system shall provide directional guidance to help users navigate.

Input Sources:

  • GPS navigation app (turn-by-turn directions)
  • User voice command: "Take me to Cafe"
  • Pre-saved routes

Guidance Methods:

  1. Haptic directional cues: Left temple vibrates for left turn, right for right turn
  2. Audio instructions: "Turn left in 20 meters"
  3. HUD arrows: Visual overlay showing direction (optional)

Features:

  • Announce street names and landmarks
  • Count down distance to next turn
  • Alert if user goes off-route
  • Suggest alternate path if obstacle blocks route

Performance:

  • Positioning accuracy: ±5m (GPS-based)
  • Turn instruction timing: 20-30m before turn
  • Re-routing time: < 5s

Validation:

  • Navigate test routes in various environments
  • Compare to Google Maps walking navigation
  • NDIS user acceptance testing

Priority: P1 (High - valuable NDIS feature)

FRD-WALK-004: Indoor Positioning (Optional)​

Description: The system shall enable indoor navigation where GPS is unavailable.

Methods:

  1. Visual SLAM: Build map using camera + IMU
  2. Wi-Fi/Bluetooth beacons: Detect known access points
  3. Dead reckoning: IMU-based step counting + heading

Accuracy:

  • ±2m within mapped building
  • ±10m with dead reckoning only

Use Cases:

  • Navigate inside shopping malls, airports, hospitals
  • Office wayfinding (enterprise variants)

Priority: P2 (Medium - future enhancement)

9. Fall Detection and Prevention​

9.1 Fall Detection System​

Already defined in FRD-SENS-IMU-003 (see section 3.2)

9.2 Fall Prevention​

FRD-WALK-005: Pre-Fall Warning​

Description: The system shall detect potential pre-fall conditions and warn the user.

Pre-Fall Indicators:

  1. Loss of balance: Rapid sway ( > 30° tilt in < 500ms)
  2. Stumble detection: Sudden forward acceleration + foot shuffle
  3. Gait irregularity: Step timing variance > 200ms
  4. Near-collision: Approaching obstacle with insufficient braking

Actions:

  • Immediate strong haptic pulse (both temples)
  • Audio alert: "Careful! Watch your step"
  • Brighten HUD (if dim) to increase visibility

Performance:

  • Detection latency: < 200ms
  • Warning time: 500-1000ms before likely fall
  • Accuracy: > 70% (challenging to validate)

Validation:

  • Motion capture analysis of falls
  • User testing with intentional loss-of-balance
  • Elderly user testing in controlled environment

Priority: P1 (High - preventative safety)

10. Health Monitoring System​

10.1 Continuous Health Tracking​

FRD-HEALTH-001: Heart Rate Monitoring Service​

Description: The system shall continuously monitor heart rate (when enabled) and detect anomalies.

Monitoring Mode:

  • On-demand: User initiates measurement
  • Continuous: Background monitoring every 5-10 minutes
  • Workout mode: Real-time during exercise

Anomaly Detection:

  • Tachycardia: Resting HR > 100 BPM for > 2 minutes
  • Bradycardia: Resting HR < 50 BPM for > 2 minutes
  • Irregular rhythm: HRV variability > 3 standard deviations

Actions:

  • Log HR data to health history
  • Alert user if anomaly detected
  • Suggest medical consultation (not diagnostic advice)
  • Sync to companion app for trend analysis

Privacy:

  • Health data stored locally, encrypted
  • Cloud sync opt-in only
  • No third-party sharing without explicit consent

Priority: P1 (High for health variants)

FRD-HEALTH-002: Activity & Fitness Tracking​

Description: The system shall track physical activity metrics.

Metrics:

  • Step count (IMU-based pedometer)
  • Distance traveled (step length calibration)
  • Calories burned (HR-based + activity level)
  • Active minutes (HR > 100 BPM)
  • Workout detection (running, cycling, gym)

Outputs:

  • Daily summary: steps, distance, calories, active minutes
  • Weekly/monthly trends
  • Goal progress (user-set goals)

Integration:

  • Export to Apple Health, Google Fit
  • Sync with GF-NF Fitness Edition AI coach

Priority: P1 (High for fitness variants)

FRD-HEALTH-003: Stress & Wellness Monitoring​

Description: The system shall detect stress levels and recommend wellness actions.

Stress Indicators:

  1. Elevated resting heart rate
  2. Reduced heart rate variability (HRV)
  3. Shallow breathing (detected via voice mic or HR patterns)
  4. Prolonged activity without rest

Stress Levels:

  • Low: HRV normal, HR normal
  • Moderate: HRV 10-25% below baseline
  • High: HRV > 25% below baseline + elevated HR

Actions:

  • Low: No action
  • Moderate: "You seem stressed. Take a few deep breaths?"
  • High: "High stress detected. Consider taking a break. Breathing exercise?"

Wellness Features:

  • Guided breathing exercises (audio + HUD visualization)
  • Meditation mode (dim HUD, play calming audio)
  • Activity break reminders (every 2 hours)

Priority: P2 (Medium - wellness feature)

11. Safety Intelligence Engine​

11.1 Multi-Layer Safety System​

FRD-SAFETY-001: Layered Safety Architecture​

Description: The system shall implement a multi-layer safety monitoring architecture.

Safety Layers:

Layer 1: Physical Safety

  • Battery thermal monitoring
  • Device surface temperature
  • Impact/drop detection
  • Fall detection

Layer 2: Environmental Safety

  • Hazardous gas detection
  • Heat stress monitoring
  • Poor air quality alerts

Layer 3: User Health Safety

  • Heart rate anomalies
  • Fatigue detection
  • Stress monitoring

Layer 4: Situational Safety

  • Walking assist (obstacles, hazards)
  • Driving mode (distraction prevention)
  • High-risk environment detection (loud noise, darkness)

Coordination:

  • Safety engine prioritizes alerts by severity
  • Multiple simultaneous alerts are combined intelligently
  • User can configure alert sensitivity (normal, high, max)

Priority: P0 (Critical - safety framework)

FRD-SAFETY-002: Emergency Response System​

Description: The system shall support emergency response features.

Emergency Triggers:

  1. Fall detection with no user response (30s timeout)
  2. User manual SOS (voice: "Call emergency" or button press)
  3. Critical health event (HR < 40 or > 180 BPM for > 1 min)
  4. Severe environmental hazard (VOC spike)

Emergency Actions:

  1. Send SOS alert to emergency contacts (via companion app)
  2. Include GPS location, timestamp, sensor data
  3. Audio message: "Emergency alert sent to [contacts]"
  4. Optional: Auto-dial emergency services (with user pre-authorization)

Emergency Contact Configuration:

  • Store 1-5 emergency contacts
  • Include relationship, phone number, email
  • Test alert system regularly

Privacy:

  • Emergency data logging separate from normal use
  • Cannot be disabled in NDIS/assistive variants
  • Can be disabled in consumer variants (with liability waiver)

Priority: P0 (Critical - emergency safety)

FRD-SAFETY-003: Safety Mode Profiles​

Description: The system shall support different safety configuration profiles.

Profiles:

Minimal (Consumer Default):

  • Basic fall detection
  • Walking assist (if enabled)
  • No automatic emergency alerts

Standard (NDIS Default):

  • Full fall detection + emergency alerts
  • Walking assist always enabled
  • Health monitoring + anomaly alerts
  • Fatigue warnings

Maximum (High-Risk Users):

  • All safety features enabled
  • Aggressive alert thresholds
  • Mandatory emergency contacts
  • Cannot disable critical safety features

Custom:

  • User configures individual feature sensitivity
  • Saved as named profiles (e.g., "Work", "Hiking", "Home")

Priority: P1 (High - user customization)

12. Calibration and Maintenance​

12.1 Factory Calibration​

FRD-CAL-001: Factory Sensor Calibration​

Description: All sensors shall be calibrated during manufacturing.

Calibration Procedures:

IMU Calibration:

  1. Accelerometer: 6-position calibration (each axis ±1g)
  2. Gyroscope: Zero-motion bias calibration
  3. Magnetometer: Figure-8 motion in 3D space

ToF Calibration:

  • Offset calibration vs. known distances (0.5m, 1m, 2m)
  • Cross-talk compensation (8×8 zones)

LiDAR Calibration:

  • Factory pre-calibrated (no field calibration needed)

Health Sensors Calibration:

  • MAX30102: LED current optimization
  • MLX90614: Factory calibrated (no field adjustment)

Calibration Storage:

  • Store calibration coefficients in secure EEPROM
  • Include calibration date, serial number
  • Include pass/fail status

Validation:

  • 100% of units must pass calibration
  • Re-test after firmware updates that touch sensor drivers

Priority: P0 (Critical - manufacturing)

FRD-CAL-002: Field Calibration​

Description: Users shall be able to re-calibrate sensors in the field.

User-Initiated Calibration:

  1. Access via Settings > Sensors > Calibration
  2. Follow on-screen guided instructions
  3. Complete calibration in < 2 minutes
  4. System provides pass/fail feedback

Automatic Calibration:

  • IMU bias updates during stationary periods
  • Magnetometer soft-iron calibration during normal use
  • Health sensor baseline adaptation (first 7 days)

Calibration Indicators:

  • "Calibration needed" notification if accuracy degrades
  • Last calibration date shown in settings
  • Calibration quality score (good, fair, poor)

Priority: P1 (High - user experience)

12.2 Sensor Health Monitoring​

FRD-CAL-003: Sensor Diagnostics​

Description: The system shall monitor sensor health and detect failures.

Health Checks:

  1. IMU: Check data continuity, noise levels, bias drift
  2. ToF: Check signal quality, valid range data
  3. LiDAR: Check measurement consistency
  4. BME688: Check plausible value ranges
  5. MAX30102: Check LED functionality, signal amplitude

Failure Detection:

  • Sensor not responding (I2C timeout)
  • Sensor returning invalid data (out of range, all zeros)
  • Sensor noise exceeding threshold

Actions:

  • Log sensor errors to diagnostic log
  • Alert user if critical sensor fails (e.g., IMU for walking assist)
  • Attempt sensor reset and retry
  • Graceful degradation (disable features that depend on failed sensor)

Diagnostic Mode:

  • Accessible via Settings > Diagnostics
  • Shows real-time sensor data streams
  • Allows manual sensor testing
  • Provides diagnostic report for support

Priority: P1 (High - reliability)

13. Integration with Other Systems​

13.1 AI System Integration​

FRD-INT-001: Sensor Data to AI Context​

Description: Sensor data shall be provided to the AI system as context for intelligent responses.

Context Provided:

  • Current activity state (walking, running, stationary, etc.)
  • Environmental conditions (temperature, humidity, indoor/outdoor)
  • Health state (heart rate, fatigue level, stress level)
  • Safety alerts (active hazards, fall detected, etc.)
  • User location and movement history

Use Cases:

  • AI adapts response length: Short responses when walking, long when stationary
  • AI suggests breaks when fatigue detected
  • AI adjusts voice volume based on ambient noise (not in this FRD, but related)
  • AI provides location-specific information (weather, nearby services)

Privacy:

  • Sensor context aggregated and anonymized
  • No raw sensor data sent to cloud without consent
  • User can disable sensor context sharing to AI

Priority: P1 (High - AI integration)

13.2 HUD Integration​

FRD-INT-002: Sensor Data Visualization​

Description: Key sensor data shall be visualizable on the HUD.

Visualizations:

Walking Assist Mode:

  • Ground profile line (green = safe, yellow = caution, red = danger)
  • Obstacle proximity indicators (expanding circles)
  • Direction arrows for navigation

Health Dashboard (quick glance):

  • Heart rate icon + BPM number
  • Step count for today
  • Activity ring (progress toward goal)

Environmental Info:

  • Temperature, humidity (small icons)
  • AQI color indicator (green, yellow, red)

Fall Detection:

  • Large red "FALL DETECTED" alert overlay
  • "Are you okay?" prompt with Yes/No

HUD Design Principles:

  • Minimal, non-intrusive by default
  • High contrast, readable in all lighting
  • Auto-hide when not needed
  • User can customize visibility

Priority: P1 (High - UX integration)

13.3 Companion App Integration​

FRD-INT-003: Sensor Data Sync​

Description: Sensor data shall sync to companion app for analysis and history.

Synced Data:

  • Daily health metrics (HR, steps, calories, sleep)
  • Fall detection events (timestamp, location, sensor data)
  • Obstacle detection logs (for debugging)
  • Environmental exposure history
  • Sensor calibration status

Sync Methods:

  • Real-time via Bluetooth (when connected)
  • Batch sync via Wi-Fi (daily summary)
  • Manual sync on-demand

App Features:

  • View trends and charts
  • Export health data (CSV, JSON)
  • Share reports with doctor (PDF)
  • Configure sensor settings remotely

Priority: P1 (High - ecosystem integration)

14. Performance Requirements​

14.1 System Performance​

MetricRequirementValidation Method
Sensor sampling latency< 5ms (IMU), < 50ms (others)Timestamp analysis
Fusion engine update rate100 Hz (10ms cycle)Real-time profiling
Walking assist alert latency< 300ms (detection to haptic)Instrumented testing
Fall detection response time< 2s (fall to alert)Drop test analysis
Power consumption< 500mW (all sensors active)Power measurement
Thermal load< 0.5W heat dissipationThermal imaging
Calibration time< 2 minutes (user-guided)User timing study

14.2 Reliability Requirements​

MetricRequirementValidation Method
Sensor failure rate< 0.1% over 2 yearsField reliability testing
Fall detection false negatives< 10% (must not miss real falls)Controlled fall testing
Fall detection false positives< 5% (must not cry wolf)Long-term user testing
Walking assist false positives< 8% (can't over-alert)Obstacle course testing
System uptime> 99.9% (excluding crashes)Field telemetry

14.3 Safety Requirements​

MetricRequirementValidation Method
Emergency alert delivery> 99% success rateNetwork testing
Critical alert latency< 1s (hazard to user notification)End-to-end test
Laser eye safetyIEC 60825-1 Class 1 complianceCertified lab testing
Skin contact safetyMax temp 38°C (sensor contact points)Thermal testing
Privacy complianceGDPR, HIPAA (health data) compliantLegal audit

15. Testing and Validation​

15.1 Unit Testing​

Per-Sensor Tests:

  • IMU: Orientation accuracy, gyro drift, mag calibration
  • ToF: Distance accuracy across range, multi-zone correctness
  • LiDAR: Distance accuracy, eye safety, response time
  • BME688: Measurement accuracy vs. reference chamber
  • MAX30102: HR accuracy vs. medical device, SpO₂ accuracy
  • MLX90614: Temperature accuracy vs. clinical thermometer

Test Environment:

  • Controlled lab conditions
  • Calibrated reference equipment
  • Temperature chamber testing
  • 100% sensor coverage

Pass Criteria:

  • All sensors meet datasheet specifications
  • Calibration improves accuracy by > 20%
  • Zero eye safety violations

15.2 Integration Testing​

Sensor Fusion Tests:

  • 6DoF pose accuracy (IMU + mag + camera)
  • Obstacle map correctness (ToF + LiDAR + IMU)
  • Drift compensation effectiveness

Safety System Tests:

  • Fall detection with crash test dummies
  • Walking assist obstacle course
  • Multi-hazard scenario handling

AI Integration Tests:

  • Context-aware AI responses
  • Sensor data feeding into AI reasoning
  • Privacy enforcement (no leaks)

Pass Criteria:

  • Fusion accuracy meets requirements
  • Safety systems trigger correctly
  • AI receives correct context data

15.3 Field Testing​

Real-World Testing:

  • NDIS participants (50+ users, 3+ months)
  • Enterprise workers (20+ users, 1+ month)
  • Fitness athletes (30+ users, 2+ months)

Test Environments:

  • Indoor, outdoor, various lighting
  • Different weather conditions
  • Urban, suburban, rural settings

Metrics:

  • User satisfaction surveys
  • Safety incident reports
  • Feature usage analytics
  • Sensor failure logs

Pass Criteria:

  • User satisfaction > 4.5/5
  • Zero safety incidents caused by system failure
  • < 5% sensor failure rate

15.4 Compliance Testing​

Regulatory Tests:

  • Laser safety (IEC 60825-1)
  • Medical device compliance (if applicable)
  • Privacy compliance (GDPR, HIPAA)
  • EMC testing (sensor immunity)

Pass Criteria:

  • All certifications obtained before launch

16. Traceability Matrix​

FRD RequirementSystem SRSHardware ReqTest Case
FRD-SENS-IMU-001REQ-SENS-001REQ-HW-120TC-IMU-001
FRD-SENS-IMU-002REQ-SENS-002REQ-HW-120TC-GESTURE-001
FRD-SENS-IMU-003REQ-SENS-003REQ-HW-120TC-FALL-001
FRD-SENS-IMU-004REQ-SENS-004REQ-HW-120TC-ACTIVITY-001
FRD-SENS-TOF-001REQ-SENS-010REQ-HW-121TC-TOF-001
FRD-SENS-TOF-002REQ-SENS-011REQ-HW-121TC-GESTURE-002
FRD-SENS-LIDAR-001REQ-SENS-020REQ-HW-122TC-LIDAR-001
FRD-SENS-LIDAR-002REQ-SENS-021REQ-HW-122TC-HAZARD-001
FRD-SENS-ENV-001REQ-SENS-030REQ-HW-123TC-ENV-001
FRD-SENS-HEALTH-001REQ-SENS-040REQ-HW-124TC-HR-001
FRD-SENS-HEALTH-002REQ-SENS-041REQ-HW-124TC-SPO2-001
FRD-WALK-001REQ-WALK-001MultipleTC-WALK-001
FRD-WALK-002REQ-WALK-002REQ-HW-122TC-WALK-002
FRD-SAFETY-001REQ-SAFE-001MultipleTC-SAFE-001
FRD-SAFETY-002REQ-SAFE-002MultipleTC-EMERG-001

(Full matrix with 50+ requirements maintained separately)

17. Glossary​

TermDefinition
IMUInertial Measurement Unit (accelerometer + gyroscope + magnetometer)
ToFTime-of-Flight (distance measurement via light pulse timing)
LiDARLight Detection and Ranging (laser-based distance measurement)
PPGPhotoplethysmography (optical heart rate measurement)
SpO₂Blood oxygen saturation percentage
HRVHeart Rate Variability (measure of heart rhythm variation)
VOCVolatile Organic Compounds (gases)
6DoFSix Degrees of Freedom (3D position + 3D orientation)
SLAMSimultaneous Localization and Mapping
HALHardware Abstraction Layer
EKFExtended Kalman Filter (sensor fusion algorithm)

Document Approval​

Approved by:

  • Hardware Lead: _________________ Date: _______
  • Safety Engineer: _________________ Date: _______
  • Software Architect: _________________ Date: _______
  • Product Manager: _________________ Date: _______
  • NDIS Compliance Officer: _________________ Date: _______

END OF FRD: SENSOR & SAFETY SYSTEMS

This FRD defines the complete sensor suite and safety intelligence for GROOT FORCE. These systems are mission-critical for our assistive and industrial variants, enabling users to navigate safely, detect health issues, and respond to emergencies. Every sensor, every algorithm, every alert is designed with one goal: keep the user safe.