Research On The Subtle Gestures Recognition Algorithm Based On EMG And FSR And Its Application

With the development of electronic technology and computer science,traditional input methods such as mouse and keyboard are unable to meet people's requirements.Recently,gesture input has been an important means of human-computer interaction due to the natural and intuitive of gestures.Human-computer interaction based on gestures emphasized more on the user-centered,and it is more suitable for users' habits,which provides a natural and effective human-computer interaction.Since the current gestures recognition algorithm involves few kinds of subtle gestures,and the recognition mainly focuses on rough gestures which can't achieves the natural interaction.So the research of subtle gestures recognition has great significance for natural human-computer interaction.There are little research on the subtle hand gestures,and the recognition methods based on single signal are limited to itself.For example,the recognition based on computer vision is easily affected by light and occlusion between fingers,and the recognition based on inertial signals can only detect the movements of gestures and other rough movements.So in this paper,we propose a subtle gestures recognition algorithm based on EMG and FSR,and the algorithm is verified by the application based on the subtle gestures recognition.The mainly work of this paper are as follows:1.We build a wearable gesture signals acquisition system to complete the collection and transmission of EMG signals and press signals.The system uses MYO armband to acquire the forearm EMG signals,and the system uses the force sensitive resistors to acquire the muscle expansion pressure signals of the back of the hand.The FSR is fixed on the glove,and the acquisition circuit is integrated into a collection watch.The collected signals are sent to the computer via Bluetooth.2.We propose an algorithm based on EMG and FSR to achieve the recognition of subtle gestures.The algorithm has an activity detection on the collected signals,and then the feature extraction is performed.The algorithm selects the top 50 best features of a gesture.After that we compare the recognition rate of different machine learning algorithms,and the machine learning algorithm with the highest recognition rate is selected to construct the classifier.Finally,the classifier is used for gesture classification in real time.3.We verify the performance of the algorithm via the writing system based on user gestures.The result of subtle gestures recognition is used to detect the pen-holding gestures,and Wii Remote is used to track the pen.Different pen-holding gestures caused different writing effects,and the system saves it in the form of electronic document.Finally we have a research and discussion on the potential application of subtle gestures recognition based on EMG and FSR.Finally the FSR glove,the collect watch and MYO armband forms a wearable device.The algorithm based on EMG and FSR can identify 21 pre-defined gestures,and the recognition accuracy is 96.05%,which is higher than use EMG only(81.7%)and use FSR only(79.52%).The human-computer interaction application has verified the performance of the algorithm and the wearable device.The wearable gesture recognition device makes the human-computer interaction more natural,and it overcomes the limitation of the natural environment.The algorithm can be applied to the unmanned aerial vehicle control,robot control and so on.