FIELD: vehicles.
SUBSTANCE: in the method for detecting and verifying the reliability of abnormal impacts on a vehicle, data related to parameters of vehicle movement is obtained, data, which is an abnormal impact on a vehicle, is recorded and transmitted to a remote station. A correlation filter based on neural network is applied to databases of abnormal impacts of true and false types. On the stage of application, data, which is an abnormal impact, is compared with data stored in the specified databases. If abnormal impacts correspond to the true type, the abnormal impact is verified to be a true event. If abnormal impacts correspond to the false type, the abnormal impact is verified to be a false event.
EFFECT: accuracy and reliability are increased.
15 cl, 2 dwg
