Design And Implementation Of Moving Hand Gestures Recognition System Based On MEMS Sensors

Looking into current existing gesture recognition technology,it mostly relies on external devices,such as cameras.Because of this dependence the recognition system is quite restricted by environmental condition,like low light or shadow which may cause recognition hardly to be carried out or error.Environmental dependence gesture recognition due to additional equipment required,makes this technology expensive and not reliable therefore it is difficult to have wide applications.Due to these existing problems,it is considered to study a feasibility of applying MEMS sensors in gesture acquisition based on MEMS sensor’ s characteristics of small size,light weight,fast response,to achieve a convenient,low cost,highly reliable and widely applicable gesture recognition technology solution.The thesis is subjected to solve the main 2 problems,one is gesture signals acquisition,and another is gesture transmission and its transmission rate matching among modules in order to realize a dynamic,real-time gesture recognition system.Design of this gesture recognition system mainly covers signal collection,transmission,data verification and processing,data analysis and recognition.Gesture signal collection is designed based on deep understanding of hands motion characteristics to identify key motion nodes where there is distribution of total 28(14 on single hand)MEMS sensor in order to acquire node movement data.Each MEMS sensor contains 2 dimensions of node movement data,acceleration and gyroscope signal to record node movement track data when gesture is implemented.Data segmentation,read and write of the 14 sensors is conducted through 8 channel I2C protocol and then integrated into the microprocessor to formulate hand movement data before being transmitted to PCs or smart phones through bluetooth protocol.In PC or smart phones the data is verified,identified and finally transformed into text or audio message for people’ s recognition.Lastly it is to verify the system by testing,including functionality and reliability test.Analysis of data frame loss rate statistics as well as test results comparison is conducted to verify cause effect relationship.Tests are conducted under different conditions to verify gesture recognition accuracy rate and stability,including Bluetooth compatibility with different brands of PCs or smart phones when the gesture acquisition system connecting with PCs or smart phones through wireless Bluetooth.The experimental results show that the gesture recognition system based on MEMS sensor meets the design expectation,and the system achieves the design target.