User-customizable machine-learning in radar-based gesture detection

A method for dynamically generating a machine-learning model for an unknown gesture, the method comprising:

• transitioning a radar-based gesture detection system into a gesture-learning mode;
• initiating, via a radar emitter of the radar-based gesture detection system, transmission of a radar field using at least one single-purpose antenna of the radar-based gesture detection system that is configured for transmission, the initiating comprising:
• initiating transmission of the radar field with a first configuration for a first iteration of analysis used to determine one or more identifying characteristics about a gesture that is undefined to the radar-based gesture detection system by selecting, as the first configuration, a first radar field configuration comprising a first field size for location detection of the gesture;
• and
• reconfiguring the radar emitter to emit the radar field with a second configuration for a second iteration of analysis used to determine the one or more identifying characteristics by selecting, as the second configuration, a second radar field configuration comprising a second field size that is different from the first field size for micro-gesture detection of the gesture;
• receiving, using at least one single-purpose antenna of the radar-based gesture detection system that is configured for reception, a first incoming radio frequency (RF) signal generated by at least a first portion of the radar field with the first configuration and a subsequent, second incoming RF signal generated by at least a second portion of the radar field with the second configuration, reflecting off the gesture;
• generating, using at least one digital signal processing component of the radar-based gesture detection system, a first digital capture of the first incoming RF signal and a second digital capture of the second incoming RF signal;
• analyzing the first digital capture in the first iteration of analysis and the second digital capture in the second iteration of analysis to learn the one or more identifying characteristics of the gesture;
• reconfiguring the radar-based gesture detection system to detect a subsequent instance of the gesture responsive to identifying the one or more identifying characteristics of the gesture from a subsequently received incoming RF signal;
• and
• transitioning the radar-based gesture detection system out of the gesture-learning mode.

A system for dynamically learning a gesture unknown to the system, the system comprising:

• a processing system;
• a gesture training application maintained in non-transitory computer-readable storage media and comprising processor-executable instructions executable by the processing system to perform a plurality of operations;
• and
• a radar-based gesture detection component implemented at least partially in hardware, and having a programmatic interface to enable communication with the gesture training application, the radar-based gesture detection component comprising:
• a radar emitter;
• an antenna;
• and
• a machine-learning component configured to operate in concert with the radar emitter and the antenna to enable the radar-based gesture detection component to:
• initiate transmission of a radar field via the radar emitter, the radar field configured to capture information about an in-the-air gesture unknown to the radar-based gesture detection component by:
• configuring the radar emitter to emit the radar field with a first configuration for a first iteration of analysis used to determine one or more identifying characteristics about the in-the-air gesture by selecting, as the first configuration, a first radar field configuration comprising a first field size for location detection of the in-the-air gesture;
• and
• reconfiguring the radar emitter to emit the radar field with a second configuration that is different from the first configuration for a second iteration of analysis used to determine the one or more identifying characteristics about the in-the-air gesture by selecting, as the second configuration, a second radar field configuration comprising a second field size that is different from the first field size for micro-gesture detection of the in-the air gesture;
• receive, via the antenna, a first incoming radio frequency (RF) signal generated by at least a first portion of the radar field with the first configuration and a second incoming RF signal generated by at least a second portion of the radar field with the second configuration reflecting off the in-the-air gesture;
• analyze the first incoming RF signal in the first iteration of analysis and the second incoming RF signal in the second iteration of analysis to determine the one or more identifying characteristics about the in-the-air gesture;
• and
• reconfigure the machine-learning component to detect a subsequent instance of the in-the-air gesture responsive to identifying the one or more identifying characteristics about the in-the-air gesture from a subsequently received incoming RF signal.