| abstract |
The invention discloses an adaptive control method of an ECAS system for anti-rollover of automobiles. First, collect offline data, build and train an improved gated recurrent neural network for predicting the time of vehicle rollover; then, collect vehicle driving state data with a certain sampling period and perform filtering processing; then use a certain update period to use the pre-trained The improved gated recurrent neural network predicts the rollover time of the vehicle, and then updates the proportional coefficient of the adaptive PD controller; at the same time, calculates the vehicle lateral load transfer ratio with a certain control period, and obtains the deviation from the set value, The deviation is then input into the adaptive PD controller for control output; finally, the electromagnetic valve of the automobile ECAS system adjusts the height of the airbag according to the output of the adaptive PD controller. The model and parameters of the invention can be self-adjusted to achieve a better anti-rollover effect; at the same time, it is applicable to many car models, thereby reducing the workload of test engineers and improving the efficiency of loading ECAS systems on cars. |