Apparatus control device, apparatus, apparatus control method, and storage medium

An apparatus control method executed by at least one hardware processor, the method comprising:

• repeatedly acquiring input data based on sensor data indicating at least one of acceleration and angular velocity generated by application of an external force to an apparatus and detected by at least one sensor provided to the apparatus, to thereby acquire a plurality of pieces of input data, wherein each piece of input data comprises a predetermined number of feature components having values calculated from the sensor data;
• mapping the plurality of pieces of input data in n-dimensional feature space according to the values of the feature components of each of the plurality of pieces of input data, where n is the predetermined number of feature components constituting each of the plurality of pieces of input data, and classifying the acquired plurality of pieces of input data into a plurality of clusters in the n-dimensional feature space by an unsupervised clustering method based on a distribution status of the plurality of pieces of input data in the n-dimensional feature space;
• newly acquiring a piece of input data based on newly detected sensor data detected by said at least one sensor provided to the apparatus;
• mapping the newly acquired piece of input data to the n-dimensional feature space according to values of the feature components of the newly acquired piece of input data and determining, as a cluster to which the newly acquired input data belongs, a cluster which is closest to the newly acquired piece of input data in the n-dimensional feature space from among the plurality of clusters;
• calculating a distance in the n-dimensional feature space between the newly acquired piece of input data and a center of the cluster to which the newly acquired piece of input data belongs, and acquiring relationship data representing a relationship between the newly acquired piece of input data and the cluster to which the newly acquired piece of input data belongs, based on the calculated distance;
• and
• controlling movement of the apparatus based on the acquired relationship data.

A non-transitory computer-readable storage medium having stored thereon a program that is executable by a computer to control the computer to perform functions comprising:

• repeatedly acquiring input data based on sensor data indicating at least one of acceleration and angular velocity generated by application of an external force to an apparatus and detected by at least one sensor provided to the apparatus, to thereby acquire a plurality of pieces of input data, wherein each piece of input data comprises a predetermined number of feature components having values calculated from the sensor data,
• mapping the plurality of pieces of input data in n-dimensional feature space according to the values of the feature components of each of the plurality of pieces of input data, where n is the predetermined number of feature components constituting each of the plurality of pieces of input data, and classifying the plurality of pieces of input data into a plurality of clusters in the n-dimensional feature space by an unsupervised clustering method based on a distribution status of the plurality of pieces of input data in the n-dimensional feature space,
• newly acquiring a piece of input data based on newly detected sensor data detected by said at least one sensor provided to the apparatus;
• mapping the newly acquired piece of input data to the n-dimensional feature space according to values of the feature components of the newly acquired piece of input data and determining, as a cluster to which the newly acquired input data belongs, a cluster which is closest to the newly acquired piece of input data in the n-dimensional feature space from among the plurality of clusters;
• calculating a distance in the n-dimensional feature space between the newly acquired piece of input data and a center of the cluster to which the newly acquired piece of input data belongs, and acquiring relationship data representing a relationship between the newly acquired piece of input data and the cluster to which the newly acquired piece of input data belongs, based on the calculated distance, and
• controlling movement of the apparatus based on the acquired relationship data.