US Patent No. 10,338,541

MACHINE LEARNING TO ESTABLISH OPTIMAL FILTER FOR REMOVING EXTERNAL NOISE WITHOUT DEGRADING RESPONSIVITY


Patent No. 10,338,541
Issue Date July 02, 2019
Title Machine Learning To Establish Optimal Filter For Removing External Noise Without Degrading Responsivity
Inventorship Kunio Tsuchida, Yamanashi (JP)
Assignee FANUC CORPORATION, Yamanashi (JP)

Claim of US Patent No. 10,338,541

1. A machine learning apparatus for learning magnitude and cutoff frequency of a filter as a condition associated with the filter for filtering an analog input signal to a motor drive apparatus, the machine learning apparatus comprising:a processor configured to
observe a state variable that includes the magnitude and the cutoff frequency of the filter as a state that an action changes directly, and a noise component and noise amount of an output signal from the filter and a responsivity to the input signal as a state that an action changes indirectly, wherein the magnitude and the cutoff frequency of the filter are selected based on an action value table, and
perform learning the magnitude and the cutoff frequency of the filter as the condition associated with the filter in accordance with a training data set that includes the state variable,
wherein the filter is modified based on a result of said learning such that an action value of the action value table is optimized.