Gesture Recognition System Based On Inertial Sensors Of The Smart Bracelet

As one of the most natural human-computer interaction, gesture recognition has been an important research in human-computer interaction. Android and IOS rapidly growing popularity of smart phones, led to the rapid development of the smart phone hardware market. The rapid development of intelligent hardware and mobile Internet continues to accelerate, driven by the emergence of a variety of new ways of human-computer interaction, gesture recognition is the hottest new human-computer interaction. Currently, intelligent vehicles, wearable devices, automotive electronics, smart phones and other fields, have been used for gesture recognition as a new generation of human-computer interaction. In recent years, with the development of integrated circuit technology, semiconductor chips are getting smaller. The new semiconductor technology to promote the MEMS(Micro-Electro-Mechanical System MEMS) technology included in a new stage of development, so many previous bulky equipment capable of miniaturization and chip technology, MEMS-based intelligent hardware and wearable equipment after another. These gestures are universal and interactive product appears ready to do the hardware, software, prepare and market preparation. Like with voice interaction, gesture recognition will be a very interactive way electronic devices HCI imaginative new direction, gesture recognition will achieve a broader application space. Future, gesture recognition will gradually reduce the power consumption, size will be further reduced, equipment costs will be greatly reduced, gesture interaction will become more popular and diverse.Depending on the manner of gesture recognition, gesture recognition can be divided into two categories: the camera image recognition based on gesture recognition sensor-based gesture tracking. This design uses gesture recognition method is based on an acceleration sensor and angular velocity sensor. This design is mainly developed a smart bracelet set of gesture recognition capabilities. The market has some intelligent hardware products and product concepts using a gesture recognition sensor technology, such as part of the smart TV remote control, as part of the smart watch with automatic raising his bright screen, and flip wrist bright screen, but a single gesture recognition, recognition core algorithms have been installed on the computer, smart TV or smart phone. Existing gesture recognition research is embedded as a data acquisition hardware to send the collected data to a computer or smartphone, then on high-end CPU gesture data processing. This greatly limits the gesture recognition in human-computer interaction in the application space.The design aims to develop a smart wearable device that can collected data, recognition your hand signal.The system uses accelerometers and gyroscopes as gesture recognition data acquisition component, nRF51822 as the chip of data processing and transceiver, a low-power Bluetooth technology for wireless data-transmission. Using Android smartphone as a data-receiving terminal. Here is the main work:(1) The original gesture data collection and filtering has been finished, including the acceleration sensor data and gyro data acquisition and hybrid filter.(2) The acceleration-data and angular-velocity-data has been convergence by Kalman filter algorithm, the acceleration sensor drift error and cumulative error is smaller, Angular velocity data noise is smaller, this algorithm combines the advantages of the acceleration sensor and the gyroscope, so the gesture data was more stable and reliable.(3) Finally, using the simple recognition algorithm designed by me to finish the gesture recognition. Unlike conventional gesture recognition method, this product is the first time to use the hand gesture recognition algorithm in the low performance computing chip., the design define and identify 14 kinds of gestures. It simplifies the complex recognition algorithm, instead of using a simple, gesture recognition method based on spatial location, simplifying the complexity of the algorithm and improve the accuracy and the speed of identification.