US 11,851,096 B2
Anomaly detection using machine learning
Forooz Shahbazi Avarvand, Berlin (DE)
Assigned to Siemens Mobility, Inc., New York, NY (US)
Filed by Siemens Mobility, Inc., New York, NY (US)
Filed on Apr. 1, 2020, as Appl. No. 16/837,406.
Prior Publication US 2021/0309271 A1, Oct. 7, 2021
Int. Cl. B61L 29/30 (2006.01); G06N 20/10 (2019.01); B61L 13/04 (2006.01); B61L 29/32 (2006.01); G05B 13/02 (2006.01); B61L 29/22 (2006.01); G05B 23/02 (2006.01); G06F 11/34 (2006.01); G06F 11/07 (2006.01); G06F 18/214 (2023.01); G06F 18/2411 (2023.01); G06N 7/01 (2023.01)
CPC B61L 29/30 (2013.01) [B61L 13/04 (2013.01); B61L 13/042 (2013.01); B61L 29/226 (2013.01); B61L 29/32 (2013.01); G05B 13/027 (2013.01); G05B 23/0205 (2013.01); G06F 11/079 (2013.01); G06F 11/3452 (2013.01); G06F 18/214 (2023.01); G06F 18/2411 (2023.01); G06N 7/01 (2023.01); G06N 20/10 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving, by a processor, a training data set comprising a plurality of labelled time series of signal values from a track circuit in a grade crossing predictor system;
removing one or more non-unique values from each labelled time series of signal values in the plurality of labelled time series of signal values to create a first set of labelled time series of signal values;
extracting a plurality of features from the plurality of labelled time series of signal values, the plurality of features comprising:
a number of signal values for each labeled time series of signal values in the first set of labelled time series of signal values that are larger than a first threshold and smaller than a maximum impedance value;
a standard deviation for each labelled time series of signal values in the first set of labelled time series of signal values;
training a machine learning algorithm utilizing the plurality of features.