2018 Electronic Competition Question D-Gesture Recognition Device (Anhui University)

In view of the characteristics and needs of the gesture recognition device, and combined with the modular design concept, a gesture recognition device with the STM32F103C8T6 microcontroller minimum system as the core was designed and implemented. The statistical grouping method is used to obtain the mapping relationship between the capacitance value between the gesture detected by the capacitive sensor (FDC2214) and the metal plate and different gestures, and the gesture template is established by combining the endpoint detection (ED) and dynamic time warping (DTW) algorithms. The working process is: first set the working mode through the buttons, and then transmit the capacitance value information detected by FDC2214 to the controller. After the controller determines the detected capacitance value, it combines the gesture action characteristics and conducts training from three aspects: reference template, distortion degree and optimal path to detect the tester's gesture. Finally, the OLED display is used to display the recognition results and provide real-time voice broadcast. After experimental testing, the proposed control method effectively improved the real-time and accuracy of the recognition process and shortened the learning time.