Automatic segmentation of inertial sensor data

A method, comprising:

• receiving, from an inertial sensor, a stream of inertial sensor data corresponding to activity of a user of an electronic device that includes the inertial sensor;
• identifying a plurality of activity transitions in the inertial sensor data;
• and
• identifying a plurality of activity segments in the inertial sensor data, each activity segment corresponding to a portion of the sensor data between adjacent activity transitions, wherein identifying the activity transitions includes:
• generating, for each of a plurality of subsets of the inertial sensor data, a plurality of wavelet sub-bands by performing a wavelet transform;
• selecting, for each subset, one of the wavelet sub-bands;
• and
• generating, for each subset of the inertial sensor data, an augmented data subset by combining the selected sub-band with the subset of the inertial sensor data.

An electronic device, comprising:

• an inertial sensor configured to generate a stream of inertial sensor data based on activity of a user;
• a control circuit coupled to the inertial sensor and configured to:
• receive the stream;
• generate a plurality of wavelet sub-bands by performing a wavelet transform on the inertial sensor data;
• identify a wavelet sub-band of highest energy from the plurality of wavelet sub-bands;
• generate augmented inertial sensor data by combining the wavelet sub-band of highest energy to the inertial sensor data;
• and
• identify a transition in the activity of the user based on the augmented inertial sensor data.