Data-driven rollover stability of a fire truck
Logging real driving data from a fire truck and using it to build a digital twin of its roll behaviour.
Outcome
Peak roll cut by 60% with tuned suspension
Key outcomes
- 01Designed and 3D-printed the IMU casing
- 02Recorded real driving data at 50 Hz (accelerometer) and 1 Hz (GPS)
- 03Peak roll reduced from 4.7° to 1.9°
The problem
Heavy, tall vehicles like fire trucks can roll over in sharp manoeuvres. The project used real driving data to understand the truck's roll behaviour and how suspension changes would affect it.
What I did
I designed and 3D-printed the casing for the IMU and set up the in-cab data logger, so the team could record real driving data: acceleration at 50 Hz and GPS at 1 Hz.
The team built a Python and MATLAB pipeline for event triggering, bias removal and filtering. We derived a one-degree-of-freedom roll-plane model and implemented it in Simulink as a digital twin, computing roll angle, load transfer ratio and inner tyre load.
The result
Tuning the suspension analytically to a 1.3 Hz natural frequency and a damping ratio of 0.5 cut peak roll by 60%, from 4.7° to 1.9°.
Next project
Forward collision warning system
Hiring?