Research Of A MEMS Inertial Sensors-Based Air Input Pen

Pen-based interactive technology is the product of human beings for pursuing a morenatural way of interaction. There are two kinds of traditional pen-based inputtingdevice-handwriting board and laser pen. However, they all have their shortcomings. Laserpen is much simple, which has only page-turning function；handwriting board is relativelyexpensive and has inconvenient portable way. In this paper, we have study and developeda kind of aerial pen-inputting system based on MEMS inertial sensors, through whichusers can be arbitrary to input in the3D space, realize the virtual keyboard characterinputting, draw lines and make marks.The system including wireless air-pen which has inertial sensor, wireless RF receiverand the interface display program of PC. Among them, the wireless air-pen is used toacquire and process accelerometer and gyroscope data, then estimates the motioninformation of the air-pen and sends the results to the PC. The PC will display themovement information on the interface.The description method of the movement information which is based on the inertialsensors are usually includes displacement description method and attitude angledescription method.Considering the limited computational resources of the air-pen’s MCUand the wrist and elbow rotation characteristics of the input operation in this application,this paper choose attitude angle to describe the motion state of the air-pen. According tothe man-machine distance and the attitude angle the MCU can estimate the movement arcof the air-pen, and then get the displacement of the cursor on the screen which is relativeto the initial position.Considering the high frequency jittering noise of acceleration sensor and thephenomenon of temperature drift of gyroscope, this paper has achieved two kinds ofattitudeAngle detecting algorithm which is based on sensor fusion algorithm on the MCU.The attitude angle detecting algorithms are respectively based on the complementary filterand kalman filter. Then we have also analyzed the performance of two algorithms in theapplication. Finally, we test the two algorithms by controlling the air-pen to draw lines andcircles which are two basic operations. The results show that the two algorithms both have high real-time performance, but the accuracy of the kalman filter is superior to thecomplementary filter.