| 摘要 |
The invention relates to a method for estimating the appearance of a pressure defect in at least one of the tyres on the wheels of a motor vehicle, said estimation resulting from an analysis of the angular speeds of the wheels using a first algorithm based on an analysis of the difference in speed between each of the wheels and using a second algorithm based on a spectral analysis of the wheel speed, each of said algorithms providing, for each wheel, an under-inflation probability datum and an associated confidence factor datum according to a wheel speed range. The invention uses a decision module based on a neural network, the inputs of which are the probability and confidence factor data for each wheel, supplied by the aforementioned algorithms, said neural network having previously undergone a learning phase and outputting, for each wheel, a tyre under-inflation state detection score. |
