Electric field touchscreen

An electric field touchscreen, comprising:

• a set of one or more transmit electrodes that are configured to generate an electric field (e-field);
• a plurality of receive electrodes that are each configured to sense an e-field variance;
• a controller configured to continuously receive signal data from the plurality of receive electrodes and output a stream of digital signal data that represents e-field signal deviations;
• and
• an application processor configured to perform the following:
• receive the stream of digital signal data,
• process the received digital signal data using a machine learning model to determine a first touch event and a first location on a display screen of the touchscreen,
• process the touch event;
• determine, from the received stream of digital signal data, whether signal strength has dropped below a threshold;
• responsive to a determination that the signal strength has dropped below the threshold, perform the following:
• disable baseline calibration for the stream of digital signal data,
• process the received stream of digital signal data through a low-pass filter to generate a first filtered stream of digital signal data,
• process the received stream of digital signal data through an absolute value average baseline filter to determine an average of the absolute values of the received stream of digital signal data;
• determine a difference between the first filtered stream of digital signal data and the average of the absolute values of the received stream of digital signal data to generate a second filtered stream of digital signal data;
• process the second filtered stream of digital signal data using the machine learning model to determine a second touch event and a second location on a display screen of the touchscreen;
• and
• process the touch event.

A non-transitory machine-readable storage medium that provides instructions that, if executed by a processor of an electric field (e-field) touchscreen, will cause said processor to perform operations comprising:

• continually receiving a stream of digital signal data that represents e-field signal deviations detected from each of a plurality of receive electrodes, wherein a display screen of the e-field touchscreen includes four edges, and wherein the plurality of receive electrodes includes at least four electrodes located along the four edges of the display screen respectively;
• processing the stream of digital signal data using a machine learning model to determine a first touch event and a first location on a display screen of the touchscreen;
• processing the touch event;
• determining, from the received stream of digital signal data, that signal strength has dropped below a threshold;
• and
• responsive to the determining that signal strength has dropped below the threshold, performing the following:
• disabling baseline calibration for the stream of digital signal data;
• processing the received stream of digital signal data through a low-pass filter to generate a first filtered stream of digital signal data,
• processing the received stream of digital signal data through an absolute value average baseline filter to determine an average of the absolute values of the received stream of digital signal data,
• determining a difference between the first filtered stream of digital signal data and the average of the absolute values of the received stream of digital signal data to generate a second filtered stream of digital signal data,
• processing the second filtered stream of digital signal data using the machine learning model to determine a second touch event and a second location on the display screen of the touchscreen;
• and
• processing the second touch event.