Research And Design Of Air Handwriting Trajectory Detection System

There are many kinds of input ways ofcomputer information.Now people are more willing to use handwriting input,so it appears in all aspects of life with the continuous development of intelligent devices.The most common handwriting input way belongs to the input of two-dimensional space information that useshandwriting pen to write in a fixed handwritten plane for collecting the trajectory or data information of writing of the handwritten plane,and then transferring the data to the terminal equipment for processing.This paper mainly studies the data input in three-dimensional space.At present,the three-dimensional trajectory tracking system is widely used in the human-computer interactionof virtual reality,including magnetic tracking system,laser tracking system and based on three-dimensional computer vision tracking system.These tracking systems have high accuracy,but their areshortcomings above high cost,large volume,uneasy to carry,unsuitable for three-dimensional air handwriting trajectory detection system.In this paper,a motion trajectory detection system based on MEMS acceleration sensor is designed.The system has the advantages of low cost,small volume and easy to carry.It is in line with the study of air handwriting trajectory detection system.This paper mainly uses the high precision six-axis sensor MPU6050 to detect the integration and angular velocity information of the handwriting during the air movement.It is based on Newton's second law,through integrating the detected data to obtain a speed,and then integrating the speed to obtained displacement data.Since the direction change occurs in the handwriting process,the rotation matrix needs to be calculated by the angular velocity,the integration data can be fixed in the same coordinate system by means of coordinate transformation,and then using two the Newton's second law to calculate the displacement,and these calculated displacement points can be connected to get the air handwriting trajectory.However,in the measurement process,some changes of the external environment will bring in random errors,or the sensor itself causes the measured acceleration and angular velocity data error,or ashaking of the hand caused errors,so that there are factors leading to the final movement trajectory display incorrectness.In this paper,in order to reduce these random errors and measurement errors,the kalman filter is designed and implemented to reduce the random error.Based on the principle of zero speed compensation algorithm and multi-axis dynamic switching algorithm,this paper designs and realizes the compensation of the detected acceleration.The accumulation of errors in the integration process has been compensated.The system can be divided into two modules: data detection module and the upper computer processing module.The data detection module mainly detects the acceleration and angular velocity data of the hand movement.The upper computer processing module processedincluding coordinate transformation,filtering,data integration and compensationby the related algorithms.In the hardware implementation,the MPU6050 module is used to detect the acceleration and angular velocity information of the handwriting motion process.Then,the MPU6050 is connected with the computer through the USB to TTL level serial module,and the detected information is uploaded to the upper computer and saved as a file;In the software implementation,the software implementation of all the algorithms are processed on the MATLAB,including coordinate conversion,kalman filter,error compensation,integration and so on.In the end,the system's experiment process is depicted in this thesis,as well as the experimental results.By analyzing these results,the thesis gets the conclusion that this system's error is less than 15%,which can meet the simple trajectory recognition.In addition,based on the experiment and analyzing results,the system's improving methods future's research works have been proposed.