A Human-Computer Interaction Technique Based On Gyro Sensor

Human-Computer Interaction Technology, as a bridge of communication between people and devices, has been developed from traditional command-line interface into graphical user interface. It makes many people who didn't have professional computer training to use computer for their own services. However, the interactive interface provided by graphical user interface is still insufficient on meeting people's needs of random input, novel sensors make encouraging progress in humanization and intellectualization of human-computer interaction.Traditional interactive equipment contains mouse, keyboard, handwritten panel, touch screen, remote controller and so on. The software in these interactive devices can convert the changing state of devices into a variety of interactive semantics, and then send them to target system, such as increase the volume, change television channels, turn page in PPT etc. However, there are some shortcomings in traditional input device. Their interactive modes are still in two-dimensional stage, it can't input in 3D, only depending on a specific plane. For example, when a teacher gives a lecture using PPT, he can only do up and down operation by mouse and keyboard at the fixed position (e.g. platform).With the development of sensor technology (e.g. MEMS technology has become more mature) and the declining price of sensors, accelerometers, gyroscopes have been penetrated from professional fields (such as defense, aviation and aerospace) into other areas, so sensor technology has been similarly applied in the field of human-computer interaction .It promotes the development of Human-computer interaction from the traditional two-dimension to three-dimension.One of changes brought by the sensor technology is the new 3D input manner, with the help of handheld device, it can offer input information by gestures in the three-dimension space. This human-computer interactive style not only can relieve the shackles of space and cable for users and achieve interaction in the free space, but also can support a variety of shapes patterns and alphanumeric information. The key of 3D input is how to quickly identify the gesture, gesture recognition is the base of 3D input.Considering that the limitations of gesture recognition's application based on image and the inherent weaknesses of accelerometer in gesture recognition, Gesture recognition method based on angular gyroscope sensor is proposed in this paper. We designed a data acquisition hardware system based on MCU and gyroscope to recognize angle information of natural gestures. And then, we proposed a computationally simple and highly efficient feature extraction method of sensor data. It can overcome dependence on movement range for input gesture. We normalized actions in the different time into the same time scale by interpolation and sub-sampling. Finally, we apply support vector model obtained by SVM to recognize feature vector of gesture.Besides theory research, we also have experiments on the gesture recognition method based on three-dimensional angular gyroscope, there are about ten Arabic numerals from zero to nine. The experimental results show that the average recognition rate is 99.4% or more.