Explanation
Ruptela devices provide several groups of accelerometer parameters, and each group has its own purpose. Using the wrong group is the most common reason why impact events are missed or triggered falsely.
Raw accelerometer axes are measured in the device's own frame. Because the device can be installed in any orientation, X/Y/Z match the vehicle's directions (forward, lateral, vertical) only after calibration, and this conversion is applied inside the device logic, not in the raw IO values.
Guidelines
1. Choose the right IO parameters
Purpose | IOs to use | Not suitable |
Impact / accident detection | G Peak X/Y/Z (IO 584, 585, 586) and Shock Duration | Accelerometer AVG X/Y/Z (IO 755–757) |
Raw acceleration for your own analysis (e.g., incline) | Accelerometer X/Y/Z (IO 49, 50, 51) with the calibration matrix | – |
Calibration status | IO 905 Axl Calibration State or SMS command accinfo | – |
2. Decode signed values correctly
G Peak X/Y/Z are signed 16-bit values in mG. If a value is greater than 32767, subtract 65536. Example: 64000 → 64000 − 65536 = −1536 mG.
Some third-party platforms (e.g., Wialon) display these values as unsigned by default, which shows very high values such as 60000+ mG without any real impact. In this case, create a custom sensor that converts the value to a signed integer, or ask your platform provider to integrate these IOs as signed values.
Accelerometer X/Y/Z (IO 49–51) are signed 8-bit values: subtract 256 from values greater than 127. [Units – to be confirmed]
3. Do not assume a fixed axis meaning
There is no fixed rule such as "X is always forward". The meaning of the raw axes depends on how the device is mounted. During calibration the device determines the gravity (downward) direction and the forward direction of the vehicle. See How is the accelerometer calibrated?
4. Convert raw axes to the vehicle's orientation
If you need acceleration in the vehicle's directions (for example, to detect uphill or downhill driving), request the calibration matrix from the device with the SMS or server command accinfo. Example response:
AXL state:11 XYZo:7884 27 -368 Xc:1000 46 53 997 Yc:1000 5 -998 53 Zc:1000 998 3 -46
Each of Xc, Yc and Zc contains a scale (1000) followed by three coefficients. The vehicle-frame value is calculated as (c1·X + c2·Y + c3·Z) / 1000. For example: Vehicle X = (46·X + 53·Y + 997·Z) / 1000. [Formula and example – to be confirmed]
The calculation can be done on your server, or on the device using Lua scripting. If neither option suits your use case, contact Ruptela support to discuss your requirements.
5. Impact event was not generated – what to check
The accelerometer is calibrated: accinfo returns AXL state 11. If Eco-Drive data with the ACC source is received, the accelerometer is working.
Both impact conditions are met: the acceleration must stay above the Lower threshold for the configured duration, and the peak value must be exceeded. If either condition is not met, the event is not generated.
Thresholds are not set too high. If impacts are missed, lower the Lower threshold to the default value, or to a value between the default and your current setting. [Recommended values – to be confirmed]
The device is mounted rigidly to the vehicle frame and cannot move or vibrate.
6. Ignition detection by movement sensor – limitation
The accelerometer detects changes in acceleration, not movement as such. When the vehicle drives at a constant speed on a smooth road (e.g., a motorway), there may be no change to detect, so ignition based only on the movement sensor can switch off while driving. Movement detection does not depend on accelerometer calibration.
For more reliable ignition detection, combine conditions, for example "movement OR speed", and add the power supply voltage condition. You can also adjust movement sensitivity and the Custom ignition switch-off delay.
Expected outcome
Impact detection uses the correct IO parameters, your platform displays signed values correctly, and, if needed, you can calculate acceleration in the vehicle's own directions.
Related documentation
Keywords
accelerometer, G peak, impact detection, accident detection, crash, signed integer, calibration matrix, accinfo, movement sensor, ignition detection, harsh driving