Motion intention recognition from multi-channel surface electromyography through prediction of major muscle activation
Recognizing human motion intentions from the human is very important in human-robot interaction applications to ensure stability. Electromyography (EMG), which measures muscle’s electrical activity, has the advantage of high-speed synchronous control when it is applied to the wearable robot because it can speed up recognition of motion intention. However, the sEMG signal characteristics are easily changed due to electrode location sensitivity and subject sensitivity. Changes in sEMG signal characteristics require a heuristic calibration for each wear, and it has been a great difficulty in recognizing human intention and generating a robot control signal. This thesis proposes a torque estimation method using multi-channel sEMG through the prediction of major muscle activation and aims to apply it to a wearable robot through torque-based motion intention recognition. The proposed method consists of a pre-processing for predicting activation of the major muscle using multi-channel sEMG and a post-processing for estimating joint torque using the major muscle signal. Prediction of the major muscle activation is a stochastic signal decomposition of the sEMG signal in static contraction motion of 80% MVC force, and the sEMG signal is orthogonally transformed in an uncorrelated state. This muscle signal decomposition model was developed to distinguish the major muscle signal for electrode location variation and inter-subject problem. The joint torque estimation model using the major muscle activation signal was made with a nonlinear auto-regressive network that processes a dynamic time-series signal. To verify the proposed pre-processing model, the joint torque estimation model was used as it is when the electrode location and the subject changed and compared with the basic method without applying the major muscle activation model. For the electrode location variation, the torque estimation accuracy was increased than the basic method. For the inter-subject variation, the torque estimation accuracy was also increased than the basic method. The sEMG-based predicting the major muscle activation showed that the torque estimation accuracy was improved for the electrode location variation and inter-subject variation. The proposed major muscle activation prediction and application overcome the limitations of sEMG-based motion intention recognition, and can also be used in wearable robots and human-robot interfaces.
Development and Analysis of a soft Surface Electromyography(sEMG) Sensor
The number of research to develop surface electromyography (sEMG) sensor for measuring muscle activation level and its applications have increased, with growing interest in bio-signal that can identify motion intentions. In order for sEMG sensor to be widely used in a wearable devices requiring repeated and prolonged use, it needs to be small and conformable to skin surface curvature while maintaining performance on EMG signal detection. However, current commercial dry type sEMG sensor is bulky and has complex system because of the separate type gain amplifier, and since it is made of hard material, it cannot be firmly attached to curved skin surface, which can increase the noise level variation according to the pressure. Also, the rigid structure may cause discomfort to the user when worn for a long time. We developed a flexible sEMG sensor which can deform to fit to the skin curvature with soft structure acquired by conductive fabric electrode, flexible PCB and silicone rubber case. This proposed design can increase wearability while maintaining the signal quality. By confirming that the SNR of the proposed sEMG sensor is comparable to the SNR of the commercial sensor, it can be seen that the proposed design can detect the EMG signal properly. The development of such a high SNR, compatible and flexible sEMG sensor can enhance the applicability of sEMG sensor in wearable device and improve its performance that identifies user intention. We have two version of this sensor, one made of silicone and one made of fabric. These sensors were applied to wearable robot applications. Especially fabric-based sensor was applied to the respiration muscles to capture the indicators that can be used for evaluating the health state of the wearer.