Generating upsampled signal from gyroscope data

A method comprising:

• receiving, by a processor and from a plurality of mobile devices, a plurality of vibration data sets, each vibration data set comprising vibration signals detected by a sensor included in a respective mobile device and time stamps of times at which the sensor detected the respective vibration signals;
• combining the vibration signals received from each mobile device to generate an upsampled vibration signal having a sampling frequency greater than a sampling frequency at which each sensor detected the respective vibration signals, wherein combining the vibration signals comprises interleaving the vibration signals, wherein interleaving the vibration signals comprises aligning the vibration signals received from the respective mobile devices in time, and wherein aligning the vibration signals comprises identifying, from each of the received vibration data sets, a vibration signal having a largest amplitude resulting in a plurality of peak vibration signals, each associated with each sensor, and aligning the plurality of peak vibration signals to generate the upsampled vibration signal;
• processing the upsampled vibration signal with one or more machine learning filters to generate a processed upsampled signal representing a combination of the plurality of vibration data sets received from the plurality of mobile devices;
• and
• transmitting the processed upsampled signal.

A system comprising:

• one or more processors;
• and
• a non-transitory computer-readable medium storing instructions executable by the one or more processors to perform operations comprising:
• receiving, from a plurality of mobile devices, a plurality of vibration data sets, each vibration data set comprising vibration signals detected by a sensor included in a respective mobile device and time stamps of times at which the sensor detected the respective vibration signals;
• combining the vibration signals received from each mobile device to generate an upsampled vibration signal having a sampling frequency greater than a sampling frequency at which each sensor detected the respective vibration signals, wherein combining the vibration signals comprises aligning the vibration signals received from the respective mobile devices in time, and wherein aligning the vibration signals comprises identifying, from each of the received vibration data sets, a vibration signal having a largest amplitude resulting in a plurality of peak vibration signals, each associated with each sensor, and aligning the plurality of peak vibration signals to generate the upsampled vibration signal;
• processing the upsampled vibration signal with one or more machine learning filters to generate a processed upsampled signal representing a combination of the plurality of vibration data sets received from the plurality of mobile devices;
• and
• transmitting the processed upsampled signal.