Rehabilitation Institute of Chicago

Ronald James Cotton

An EMG-based control system is disclosed. The system includes an EMG sensor that streams data to a device. A neural network executed by the device decodes muscle activity into a two-dimensional continuous control signal. A network-based calibration routine optimizes the neural network to decode natural movements of the user.

Systems and methods for electromyogram-based control using neural networks

A system for utilizing wearable, wireless-connected sensors that record both IMU and EMG muscle activity for rehabilitation with kinematic monitoring is disclosed. The system runs an efficient quaternion-based complementary filter that estimates the sensor orientation while correcting for estimate drift and constraining magnetometer estimates to only influence heading. The difference in the two sensor orientations is used to estimate the joint angle, which can be further improved with joint axis estimation. Thus, successful tracking of joint angle and muscle activity in a home environment can be accomplished with the sensors and a smartphone. Further, a system utilizing these sensors may include video-capable cameras configured for capturing a video of a pose of the user, and a game framework executable by a processor in communication with the sensors and the cameras. The system may include a game framework configured to utilize a pose trajectory and one or more joint angles determined using the sensors to provide biofeedback to a user.

Growing community of inventors

Chicago, IL, United States of America

Ronald James Cotton