Systems, articles, and methods for gesture identification in wearable electromyography devices

Patent metadata

Field	Value
Publication number	US-9389694-B2
Application number	US-201414520081-A
Country	US
Kind code	B2
Filing date	Oct 21, 2014
Priority date	Oct 22, 2013
Publication date	Jul 12, 2016
Grant date	Jul 12, 2016

Abstract

Systems, articles, and methods perform gesture identification with limited computational resources. A wearable electromyography (“EMG”) device includes multiple EMG sensors, an on-board processor, and a non-transitory processor-readable memory that stores data and/or processor-executable instructions for performing gesture identification. The wearable EMG device detects and determines features of signals when a user performs a physical gesture, and processes the features by performing a decision tree analysis. The decision tree analysis invokes a decision tree stored in the memory, where storing and executing the decision tree may be managed by limited computational resources. The outcome of the decision tree analysis is a probability vector that assigns a respective probability score to each gesture in a gesture library. The accuracy of the gesture identification may be enhanced by performing multiple iterations of the decision tree analysis across multiple time windows of the EMG signal data and combining the resulting probability vectors.

First claim

The invention claimed is:

1. A method of operating a wearable electromyography (“EMG”) device, wherein the wearable EMG device includes a set of EMG sensors and a processor communicatively coupled to the set of EMG sensors, the method comprising:
detecting muscle activity of a user of the wearable EMG device by the set of EMG sensors, wherein the muscle activity corresponds to a user-performed gesture;
in response to detecting muscle activity of the user by the set of EMG sensors, providing a set of signals from the set of EMG sensors to the processor;
determining a set of features of the set of signals by the processor;
performing a series of evaluations of at least some of the features in the set of features by the processor, wherein each evaluation in the series of evaluations includes evaluating a respective feature in the set of features by the processor, and wherein a first evaluation in the series of evaluations includes evaluating a first feature in the set of features by the processor and each subsequent evaluation in the series of evaluations is based at least in part on an outcome of a previous evaluation in the series of evaluations;
determining a respective probability score of each gesture in a gesture library by the processor based at least in part on an outcome of the series of evaluations; and
identifying the user-performed gesture by the processor based at least in part on the probability score of at least one gesture in the gesture library.

2. The method of claim 1 wherein performing a series of evaluations of at least some of the features in the set of features by the processor includes performing a decision tree analysis of the set of features by the processor.

3. The method of claim 1 wherein each evaluation in the series of evaluations includes comparing a magnitude of a respective feature in the set of features to a respective value by the processor.

4. The method of claim 1 wherein determining a set of features of the set of signals by the processor includes determining at least one feature selected from the group consisting of: an average value of a signal in the set of signals, a mean value of a signal in the set of signals, a median value of a signal in the set of signals, a mode value of a signal in the set of signals, a maximum value of a signal in the set of signals, a minimum value of a signal in the set of signals, a standard deviation of a signal in the set of signals, and/or a root mean squared (“RMS”) value of a signal in the set of signals.

5. The method of claim 1 wherein identifying the user-performed gesture by the processor based at least in part on the probability score of at least one gesture in the gesture library includes identifying, by the processor, a gesture in the gesture library that has a largest probability score.

6. The method of claim 1 wherein the wearable EMG device further includes a non-transitory processor-readable storage medium communicatively coupled to the processor, and wherein the non-transitory processor-readable storage medium stores processor-executable gesture identification instructions, and wherein:
determining a set of features of the set of signals by the processor includes executing, by the processor, the processor-executable gesture identification instructions to cause the processor to determine a set of features of the set of signals;
performing a series of evaluations of at least some of the features in the set of features by the processor includes executing, by the processor, the processor-executable gesture identification instructions to cause the processor to perform a series of evaluations of at least some of the features in the set of features;
determining a respective probability score of each gesture in a gesture library by the processor based at least in part on an outcome of the series of evaluations includes executing, by the processor, the processor-executable gesture identification instructions to cause the processor to determine a respective probability score of each gesture in a gesture library based at least in part on an outcome of the series of evaluations; and
identifying the user-performed gesture by the processor based at least in part on the probability score of at least one gesture in the gesture library includes executing, by the processor, the processor-executable gesture identification instructions to cause the processor to identify the user-performed gesture based at least in part on the probability score of at least one gesture in the gesture library.

7. The method of claim 1 wherein the wearable EMG device further includes at least one inertial sensor, and wherein the method further comprises:
detecting motion of the wearable EMG device by the at least one inertial sensor, wherein the motion corresponds to the user-performed gesture;
in response to detecting motion of the wearable EMG device by the at least one inertial sensor, providing at least one signal from the at least one inertial sensor to the processor; and
processing the at least one signal from the at least one inertial sensor by the processor, and wherein identifying the user-performed gesture by the processor based at least in part on the probability score of at least one gesture in the gesture library includes identifying the user-performed gesture by the processor based at least in part on an outcome of the processing the at least one signal from the at least one inertial sensor by the processor.

8. The method of claim 1 , further comprising:
capturing a respective time-synchronized first portion of each signal in the set of signals by the processor, and wherein:
determining a set of features of the set of signals by the processor includes determining a set of features of the time-synchronized first portions of the signals in the set of signals by the processor;
performing a series of evaluations of at least some of the features in the set of features by the processor includes performing a series of evaluations of at least some of the features of the time-synchronized first portions of the signals in the set of signals by the processor; and
determining a respective probability score of each gesture in a gesture library by the processor based at least in part on an outcome of the series of evaluations includes determining a respective first probability score of each gesture in the gesture library by the processor based at least in part on an outcome of the series of evaluations of at least some of the features of the time-synchronized first portions of the signals in the set of signals.

9. The method of claim 8 , further comprising:
capturing a respective time-synchronized second portion of each signal in the set of signals by the processor, wherein:
determining a set of features of the set of signals by the processor includes determining a set of features of the time-synchronized second portions of the signals in the set of signals by the processor;
performing a series of evaluations of at least some of the features in the set of features by the processor includes performing a series of evaluations of at least some of the features of the time-synchronized second portions of the signals in the set of signals by the processor; and
determining a respective probability score of each gesture in a gesture library by the processor based at least in part on an outcome of the series of evaluations includes determining a respective second probability score of each gesture in the gesture library by the processor based at least in part on an outcome of the series of evaluations of at least some of the features of the time-synchronized second portions of the signals in the set of signals; and

generating a respective cumulative probability score of each gesture

Assignees

• Thalmic Labs Inc

Inventors

• Ataee Pedram
• Aleem Idris S
• Bailey Matthew

Classifications

• G16H50/70

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G06F3/017 Primary

  Gesture based interaction, e.g. based on a set of recognized hand gestures (interaction based on gestures traced on a digitiser G06F3/04883) · CPC title

• A61B2562/063

  in a linear array · CPC title

• A61B5/7267

  involving training the classification device · CPC title

• A61B5/681

  Wristwatch-type devices · CPC title