In order to prevent car rollover while moving

and to make the car always in a safe driving condition and to reduce traffic accidents

an ARM9-based intelligent vehicle perception and rollover warning system is designed.Monitor the roll state of the car when driving and multi-level recursive model is used to predict the vehicle roll attitude.When the predicted roll reach to the limit conditions

trigger the alarm to remind driver to be careful and to take appropriate method to prevent vehicle rollover accidents

and further improve the vehicle active safety performance.Road experiments are designed

Matlab is used to compare the predicted values with experimental data.The results show that vehicle perception and rollover warning system can predict vehicle rollover timely and accurately

which can improve vehicle active safety.