Multiple User Interaction with Audio Devices Using Speech and Gestures

A computer program product, the computer program product comprising a tangible computer readable storage medium having program instructions embodied therewith, the program instructions executable by a device to cause the device to:

• transmit a pilot signal from the device, the pilot signal having a frequency and the device comprising a voice activated interface;
• detect motion of at least one user around the device, wherein in detecting the motion, the program instructions further cause the device to:
• receive a reflected signal of the pilot signal caused by the motion of the at least one user;
• and
• detect a shift in the reflected signal from the pilot signal, wherein in detecting the shift, the program instructions further cause the device to determine one or more differences between a waveform corresponding to the pilot signal and a waveform corresponding to the reflected signal, wherein said detecting the shift comprises (i) passing the reflected signal through a high pass filter, and (ii) processing output from the high pass filter utilizing a non-equispaced fast Fourier transform (NFFT);
• associate the detected motion with a gesture via application of at least one machine learning classification technique, comprising one or more support vector machines, to information pertaining to the detected shift, thereby identifying a shift-related signature caused by the detected motion and mapping the shift-related signature to training data associated with multiple user motions;
• and
• perform a function in response to the gesture.

A system comprising:

• a memory;
• and
• at least one processor coupled to the memory and configured for:
• transmitting a pilot signal from a device comprising a voice activated interface, the pilot signal having a frequency;
• detecting motion of at least one user around the device, wherein said detecting the motion comprises:
• receiving a reflected signal of the pilot signal caused by the motion of the at least one user;
• and
• detecting a shift in the reflected signal from the pilot signal, wherein said detecting the shift comprises determining one or more differences between a waveform corresponding to the pilot signal and a waveform corresponding to the reflected signal, wherein said detecting the shift comprises (i) passing the reflected signal through a high pass filter, and (ii) processing output from the high pass filter utilizing a non-equispaced fast Fourier transform (NFFT);
• associating the detected motion with a gesture via application of at least one machine learning classification technique, comprising one or more support vector machines, to information pertaining to the detected shift, thereby identifying a shift-related signature caused by the detected motion and mapping the shift-related signature to training data associated with multiple user motions;
• and
• performing a function in response to the gesture.

A computer-implemented method, comprising:

• transmitting a pilot signal from a device comprising a voice activated interface, the pilot signal having a frequency;
• mapping a plurality of users to respective locations around the device;
• detecting respective motions of the plurality of users around the device, wherein said detecting the motion comprises:
• receiving respective reflected signals of the pilot signal from each of the plurality of users caused by the respective motions of the plurality of users;
• and
• detecting respective shifts in the respective reflected signals from the pilot signal, wherein said detecting the respective shifts comprises determining one or more differences between a waveform corresponding to the pilot signal and waveforms corresponding to the respective reflected signals, wherein said detecting the shifts comprises (i) passing the reflected signals through a high pass filter, and (ii) processing output from the high pass filter utilizing a non-equispaced fast Fourier transform (NFFT);
• associating the detected respective motions with respective gestures via application of at least one machine learning classification technique, comprising one or more support vector machines, to information pertaining to the detected shifts, thereby identifying shift-related signatures caused by the detected motions and mapping the shift-related signatures to training data associated with multiple user motions;
• and
• performing respective functions in response to the respective gestures.