• 인터넷정보학회논문지
• /
• 제25권2호
• /
• pp.57-67
• /
• 2024

사용자가 제스처를 통해 입력을 할 수 있는 방안들 중에서 근전도(EMG, Electromyography)를 통한 제스처 인식은 근육 내 작은 전극을 통해 사용자의 움직임을 감지하고 이를 입력 방법으로 사용할 수 있는 방법이다. EMG 데이터를 통해 사용자 제스처를 분류하기 위해서는 사용자로부터 수집된 EMG Raw 데이터를 머신러닝으로 학습하여야 하는데 이를 위해서는 EMG 데이터를 전처리 과정을 통해 특징을 추출하여야 한다. EMG 특성은 IEMG(Integrated EMG), MAV(Mean Absolute Value), SSI(Simple Sqaure Integral), VAR(VARiance), RMS(Root Mean Square) 등과 같은 수식을 통해서 나타낼 수 있다. 또한, 제스처를 입력으로 사용하기 위해서는 사용자가 입력하는 데 필요한 지각, 인지, 반응에 필요한 시간을 기준으로 제스처 분류가 가능한 시간을 알아내야 한다. 이를 위해 최대 1,000ms에서 최소 100ms까지 세그먼트 사이즈를 변화시켜 특징을 추출 후 제스처 분류가 가능한 세그먼트 사이즈를 찾아낸다. 특히 데이터 학습은 overlapped segmentation 방법을 통해 데이터와 데이터 사이 간격을 줄여 학습 데이터 개수를 늘린다. 이를 통해 KNN, SVC, RF, XGBoost 4가지 머신러닝 방식을 통해 이를 학습하고 결과를 도출한다. 실험 결과 실시간으로 사용자의 제스처 입력이 가능한 최대 세그먼트 사이즈인 200ms에서 KNN, SVC, RF, XGboost 4가지 모든 모델에서 96% 이상의 정확도를 도출하였다.

• 제어로봇시스템학회논문지
• /
• 제17권10호
• /
• pp.1014-1020
• /
• 2011

This presents a motion and force estimation system of human fingers by using an Electromyography (EMG) sensor module and a data glove system to be proposed in this paper. Both EMG sensor module and data glove system are developed in such a way to minimize the number of hardware filters in acquiring the signals as well as to reduce their sizes for the wearable. Since the onset of EMG precedes the onset of actual finger movement by dozens to hundreds milliseconds, we show that it is possible to predict the pattern of finger movement before actual movement by using the suggested system. Also, we are to suggest how to estimate the grasping force of hand based on the relationship between RMS taken EMG signal and the applied load. Finally we show the effectiveness of the suggested estimation system through several experiments.

• 한국전문물리치료학회지
• /
• 제11권4호
• /
• pp.61-69
• /
• 2004

EMG is used in rehabilitation research to provide a method to infer muscle function. This paper will present an introduction to interpretation of electromyography (EMG) data for physical therapists. It is important for the physical therapist to have an understanding of the collection and reduction of raw electrical data from the muscle to allow the physical therapist to interpret findings in a research report, and improve planning of clinical research projects with respect to data collection. We will discuss factors that affect the type of EMG collected and the ways in which various common methods of data reduction will impact the findings from a study that uses EMG.

• 대한안전경영과학회지
• /
• 제8권1호
• /
• pp.43-52
• /
• 2006

The goal of this study is to investigate the ergonomic factors in designing or selecting the hand tool handle. Electromyogram (EMG) were measured for various wrist postures and handle sizes under two loading conditions. Anthropometric data were measured and the correlation with EMG measurement data were analyzed. Investigations of this study show that wrist posture should be neutral for minimum muscle tension and optimum handle size can be found by measuring the EMG measurement data. It show that hand width and EMG measurement data is greatly correlated also. This study can be a guide of designing or selecting a hand tool, but further study with large sample sizes and various groups is needed for making general conclusion.

• 대한의용생체공학회:의공학회지
• /
• 제13권1호
• /
• pp.1-8
• /
• 1992

In the field of prosthesis arm control, the pattern classification of the EMG signal is a required basis process and also the estimation of force from collected EMG data is another necessary duty. But unfortunately, what we've got is not real force but an EMG signal which contains the information of force. This is the reason why we estimate the force from the EMG data. In this paper, when we handle the EMG signal to estimate the force, spatial prewhitening process is applied from which the spatial correlation between the channels are removed. And after the orthogonal transformation which is used in the force estimation process, the transformed signal Is inputed into the probabilistic model for pattern classification. To verify the different results of the multiple channels, SNR(signal to noise ratio) function is introduced.

• 전기학회논문지
• /
• 제65권1호
• /
• pp.194-199
• /
• 2016

EMG has specific information which is related to movements according to the activities of muscles. Therefore, users can intuitively control a prosthesis. For this reason, biosignals are very useful and convenient in this kind of application. Bioimpednace also provides specific information about movements like EMG. In this study, we used both EMG and bioimpedance to classify the typical hand gestures such as hand open, hand close, no motion (rest), supination, and pronation. Nine able-bodied subjects and one amputee were used as experimental data set. The accuracy was $98{\pm}1.9%$ when 2 bio-impedance and 8 EMG channels were used together for normal subjects. The number of EMG channels affected the accuracy, but it was stable when more than 5 channels were used. For the amputee, the accuracy is higher when we use both of them than when using only EMG. Therefore, accurate and stable hand motion estimation is possible by adding bioimepedance which shows structural information and EMG together.

• Physical Therapy Rehabilitation Science
• /
• 제10권3호
• /
• pp.374-378
• /
• 2021

Objective: Bilateral movement training is an effective method for upper extremity rehabilitation of stroke. An approach to induce bilateral movement through functional electrical stimulation is attempted. The purpose of this study is to develop an EMG-triggered functional electrical stimulation device for upper extremity bilateral movement training in stroke patients and test its feasibility. Design: Feasibility and Pilot study design. Methods: We assessed muscle activation and kinematic data of the affected and unaffected upper extremities of a stroke patient during wrist flexion and extension with and without the device. Wireless EMG was used to evaluate muscle activity, and 12 3D infrared cameras were used to evaluate kinematic data. Results: We developed an EMG-triggered functional electrical stimulation device to enable bilateral arm training in stroke patients. A system for controlling functional electrical stimulation with signals received through a 2-channel EMG sensor was developed. The device consists of an EMG sensing unit, a functional electrical stimulation unit, and a control unit. There was asymmetry of movement between the two sides during wrist flexion and extension. With the device, the asymmetry was lowest at 60% of the threshold of the unaffected side. Conclusions: In this study, we developed an EMG-triggered FES device, and the pilot study result showed that the device reduces asymmetry.

• 한국정보통신학회논문지
• /
• 제10권10호
• /
• pp.1779-1785
• /
• 2006

본 논문은 근전도 패턴 인식에 의한 가상 로봇팔 제어 방식을 제안한다. 고차원의 근전도 신호를 정밀하게 분류하기 위하여 혼합형 신경 회로망 방식을 사용한다. 혼합형 신경회로망은 SOFM과 LVQ로 구성되고, 고차원의 EMG 신호를 2차원 데이터로 변환한다. 3개의 표면 전극을 이용하여 EMG 신호를 측정 한다. 제안한 혼합 시스템을 이용하여 한글 자음 6개의 수화 신호를 분류한다. 가상 로봇팔 실험을 통해서 제안한 혼합 시스템을 이용한 수신호의 EMG 패턴 인식의 유용성을 확인하였다.

• 대한의용생체공학회:의공학회지
• /
• 제27권5호
• /
• pp.237-244
• /
• 2006

This study proposes an approach to the performance improvement of EMG(Electromyogram) pattern recognition. MFCC(Mel-Frequency Cepstral Coefficients)'s approach is molded after the characteristics of the human hearing organ. While it supplies the most typical feature in frequency domain, it should be reorganized to detect the features in EMG signal. And the dynamic aspects of EMG are important for a task, such as a continuous prosthetic control or various time length EMG signal recognition, which have not been successfully mastered by the most approaches. Thus, this paper proposes reorganized MFCC and HMM-GMM, which is adaptable for the dynamic features of the signal. Moreover, it requires an analysis on the most suitable system setting fur EMG pattern recognition. To meet the requirement, this study balanced the recognition-rate against the error-rates produced by the various settings when loaming based on the EMG data for each motion.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2001년도 ICCAS
• /
• pp.84.4-84
• /
• 2001

In this paper, we present a method to classify electromyogram (EMG) signals which are utilized to be control signals for patient-responsive walker-supported system for paraplegics. Patterns of EMG signals for dierent walking motions are classied via adequate filtering, real EMG signal extraction, AR-modeling, and modified self-organizing feature map (MSOFM). More efficient signal processing is done via a data-reducing extraction algorithm. Moreover, MSOFM classifies and determines the classified results are presented for validation.