Human-machine interface using eye saccade and facial expression physiological signals to improve the maneuverability of wearable robots

This paper proposed a method using eye saccade and facial expression physiological signals to interpret human intentions as a mean to operate wearable robots. Our approach used only two electrodes placed on top of the left and right ears to identify high fidelity physiological signals. With the developed machine learning algorithms, we can achieve over 97% accuracy of the classified various human eye-facial gestures, which allows us to build up intuitive Human-Machine Interaction (HMI) strategies to let disable people operate exoskeletons or wheelchairs easily and naturally. This method also enables the design of earbud-like wearable device, which can be worn comfortably for long hours to provide ubiquitous controllability to operate external smart devices. The ultimate goal of this research is to seamlessly merge human and machine all together by just using a simple wearable device.