US 12,386,324 B2
Method and device for training a data-based time determining model for determining an opening or closing time of an injection valve using a machine learning method
Andreas Hopf, Stuttgart (DE); Erik Tonner, Mehring (DE); Frank Kowol, Knittlingen (DE); Jens-Holger Barth, Fellbach (DE); Konrad Groh, Stuttgart (DE); Matthias Woehrle, Bietigheim-Bissingen (DE); Mona Meister, Renningen (DE); and Roland Norden, Kornwestheim (DE)
Assigned to ROBERT BOSCH GMBH, Stuttgart (DE)
Appl. No. 18/001,546
Filed by Robert Bosch GmbH, Stuttgart (DE)
PCT Filed Sep. 10, 2021, PCT No. PCT/EP2021/074944
§ 371(c)(1), (2) Date Dec. 12, 2022,
PCT Pub. No. WO2022/053620, PCT Pub. Date Mar. 17, 2022.
Claims priority of application No. 10 2020 211 419.6 (DE), filed on Sep. 11, 2020.
Prior Publication US 2023/0229121 A1, Jul. 20, 2023
Int. Cl. G05B 13/04 (2006.01); F02D 41/14 (2006.01)
CPC G05B 13/048 (2013.01) [F02D 41/14 (2013.01)] 10 Claims
OG exemplary drawing
 
1. A computer-implemented method for training a data-based time determining model for determining an opening or closing time of an injection valve based on a sensor signal, the method comprising the following steps:
providing at least one unlabeled analysis point time series by sampling the sensor signal of a sensor of the injection valve; and
training the data-based time determining model in order to assign a time specification, which represents a specific opening or closing duration, to an analysis point time series, wherein the training process is carried out using a first shifting function for time-shifting the analysis point time series and a second shifting function for time-shifting the time specification, wherein a consistency loss function is used for the training process, the consistency loss function indicating a difference between a first time specification, which corresponds to a model output of the time determining model for the analysis point time series time-shifted according to the first shifting function, and a second time specification, which specifies a time specification which is time-shifted according to the second shifting function and corresponds to a model output of the time determining model for the at least one analysis point time series.