Sensor Fusion Visualizer
Run the Sensor Fusion Visualizer Fullscreen
Edit the Sensor Fusion Visualizer with the p5.js editor
About This MicroSim
This visualization demonstrates sensor fusion - combining data from multiple sensors to achieve better accuracy and robustness than any single sensor alone. It shows GPS (noisy but absolute) and IMU (smooth but drifting) sensors being fused via Kalman filtering.
Embedding
You can embed this MicroSim in your website using:
1 2 | |
Features
- True Position: Blue circle follows a circular path with variations
- GPS Measurements: Cyan markers appear periodically (low rate, noisy, absolute)
- IMU Dead Reckoning: Orange trail showing integration drift
- Fused Estimate: Green trail with uncertainty ellipse (best of both)
- RMSE Comparison: Real-time error statistics for each method
Key Concepts
Why Fuse Sensors?
| Sensor | Strengths | Weaknesses |
|---|---|---|
| GPS | Absolute position, global | Low rate (~10Hz), poor in tunnels/canyons |
| IMU | High rate (~400Hz), works everywhere | Drift accumulates over time |
The Fusion Principle
Kalman filter fusion exploits complementary characteristics: - GPS corrects IMU drift with absolute measurements - IMU fills gaps between GPS updates with smooth predictions - Result: High-rate, accurate position with bounded error
Error Growth
- GPS Only: Error bounded but noisy (no smooth trajectory)
- IMU Only: Error grows unboundedly over time (drift)
- Fused: Error bounded AND smooth (best of both worlds)
Lesson Plan
Learning Objectives
- Understand why multiple sensors outperform single sensors
- Recognize drift vs noise characteristics
- Analyze fusion improvement quantitatively via RMSE
Activities
- GPS Only: Disable IMU, observe jumpy but bounded tracking
- IMU Only: Disable GPS, watch trajectory drift away
- Fusion Benefit: Enable both, compare RMSE values
- Noise Tradeoffs: Increase GPS noise, see fusion degrade gracefully
Assessment Questions
- Why does IMU-only tracking drift while GPS-only stays bounded?
- How does the fused estimate achieve lower RMSE than either sensor alone?
- What happens to fusion quality when GPS updates become less frequent?
References
- Sensor Fusion and Tracking (MATLAB)
- GPS/IMU Integration
- Chapter 15: Autonomous Systems and Sensor Fusion