BlackBox
A phone-based car black box: live telemetry, event detection and CSV export.
The idea
Turn any Android phone into a car "black box". Using the sensors every phone already carries — accelerometer, gyroscope and GPS — the app records a drive the way a dashcam or flight recorder would, then exports the raw data for analysis.
The interface borrows the feel of a modern car dashboard: a few large, glanceable cards, clear Start / Stop / Share buttons colour-coded for safety, and a red REC indicator so it's always obvious when data is being logged.
Features
- Live readouts: speed (km/h), longitudinal G-force and yaw rate.
- Performance summary: top speed, 0–100 and 100–0 km/h, distance.
- Automatic detection of hard braking, hard acceleration and turns.
- Full CSV log of every sample, shareable to any app.
- Runtime location permission handled inline on Start.
- Dashboard-style UI with cards, a pulsing REC dot and animations.
How it works
The app fuses three data sources. Android's SensorManager streams linear acceleration (converted from m/s² to g for readability) and the gyroscope's yaw rate; Google's Fused Location Provider gives speed, position and distance.
Events are flagged with thresholds plus a short dwell time, so a single noisy spike doesn't trigger a false positive — for example, braking harder than −3 m/s² held for 300 ms.
timestamp,lat,lon,speed_kmh,g_long,yaw,distance_m,event 1731612000123,46.2044,6.1432,52.3,-0.34,0.08,812.0, 1731612000345,46.2044,6.1432,41.1,-3.20,0.05,815.0,BRAKE 1731612000720,46.2045,6.1433,0.0,0.02,0.61,815.0,TURN_Droite
Every sample is appended to a CSV; events (BRAKE, ACCEL, TURN) are tagged inline, ready for Excel or Python.
A look at the app
Where these sensors are used
The same signals power real safety systems: a sudden negative G is how phones and cars detect crashes and fire airbags or emergency calls; yaw-rate sensors feed ESP / stability control to catch skids; and GPS combined with acceleration profiles drives fleet management and usage-based insurance.
Results
A working, self-contained Android app. Recording, event tagging and CSV export all run end to end, ready to open in Excel or Python. The 0–100 and 100–0 timings are GPS approximations rather than lab figures, and the detection thresholds are tuned by hand — honest trade-offs for a learning project.