| CPC B61K 9/12 (2013.01) [B61B 3/02 (2013.01); H01L 21/67733 (2013.01)] | 4 Claims |
|
1. A rail-guided cart system comprising:
a rail;
a cart to travel along the rail;
a measurement sensor fixed with respect to the rail to measure at least one of vibration and sound;
a diagnosis processor configured or programmed to diagnose a condition of the cart passing through a measurement point on the rail corresponding to a location of the measurement sensor, based on measurement data measured by the measurement sensor;
a feature extractor to extract a for-training feature included in for-training partial data which are partial data corresponding to data when the cart passes through the measurement point, from measurement data measured by the measurement sensor;
a learning controller configured or programmed to construct a learning model for diagnosis of the condition of the cart by using a for-training data set which is a set of the for-training features based on measurement data measured by the measurement sensor for a plurality of carts containing no abnormality;
a for-training partial data extractor to extract the for-training partial data from measurement data measured by the measurement sensor; and
a position sensor fixed with respect to the rail to detect the cart passing through the measurement point; and
a notifier electrically connected to the diagnosis processor to notify when the condition of the cart is abnormal; wherein
the for-training partial data are extracted from the measurement data based on a timing at which the position sensor detects the cart;
a correlation between reference data as a basis and measurement data to be extracted is determined while the measurement data are variously offset in a time axis direction; and
the partial data are extracted by cutting out from the measurement data according to a time range defined with reference to the reference data, in a state where the measurement data are offset in the time axis direction so that the correlation is largest.
|