Method for training a mapping for outputting an abnormality determination variable for a specific component mounted on a vehicle
A training method includes: a simulation step outputting, by a simulator, a characteristic variable based on an input parameter set, the parameter set being input to the simulator and indicating that a specific component is presumed to have an abnormality in advance; and a training step updating, by a training device, mapping based on an input training data and an input teaching data, the input variables being input to the training device as the training data, the input variables including the characteristic variable output in the simulation step, the abnormality determination variable being input to the training device as the teaching data, the abnormality determination variable indicating that the specific component has the abnormality.
1. A training method for training a mapping that outputs, as an output variable, an abnormality determination variable that (i) indicates an occurrence possibility of an abnormality of a specific component mounted on a vehicle and (ii) is used in a determination process for determining whether or not the specific component has the abnormality, the method comprising:
outputting, by a simulator and based on an input parameter set to the simulator that has a plurality of condition parameters related to conditions of the vehicle, characteristic variables at least one of which indicates a decrease rate of at least one of (i) vibration of the vehicle per unit time and (ii) noise of the vehicle per unit time,
the simulator storing model data that defines a vehicle model, which is a model of structure of the vehicle, and outputting the characteristic variables by inputting the input parameter set to the model data,
one of the condition parameters being a specific parameter indicating a property of the specific component, and
the input parameter set (i) indicating that the specific component is presumed to have the abnormality in advance by (ii) being a parameter set having the specific parameter that falls outside of a predetermined normal range; and
updating, by a processor, the mapping based on training data and teaching data,
the training data including a plurality of input variables for stored mapping data defining the mapping, the input variables including the characteristic variables output by the simulator, and
the teaching data including the abnormality determination variable indicating that the specific component has the abnormality, wherein:
the decrease rate is calculated (i) by subtracting an amplitude of the vibration or noise at an end of a duration of the amplitude or noise from an amplitude of the vibration or noise at a start of the duration and (ii) by dividing a value obtained through the subtraction by the duration;
the specific component is a damper comprising a damper body, which includes an elastic member, and an attenuator; and
the characteristic variables based upon which the mapping is updated include (i) a frequency of the damper at a maximum amplitude of vibration that is caused by the damper and is related to an abnormality of the damper body, (ii) a duration of vibration of the damper that is related to an abnormality of the attenuator, and (iii) a decrease rate of the vibration that is related to the abnormality of the attenuator.
2. The method according to claim 1 , wherein one of the condition parameters is a vehicle speed.
3. The method according to claim 1 , wherein one of the condition parameters is an operation amount of an accelerator pedal.
4. The method according to claim 1 , wherein the structure of the vehicle modeled by the vehicle model includes a driving wheel, a suspension, a frame, a mount, and a drive system of the vehicle.