The Research And Application Of Integrated Navigation Technology For Two Sheels Self-balancing Robot

Along with the development of science,the precision and reliability of navigation system for motor vehicle have became higher and higher,how to get a high precision navigation information has become the hot topic in the study of mobile robot navigation.This paper mainly adopts integrated navigation and data fusion technology to improve the precision of navigation and speed of attitude convergence under those condition,such as,the non-gravity disturbance of accelerometer,start-up operation and abnormal model interference problems,and use MATLAB simulation and experimental to verificate the effectiveness of relevent navigation algorithm.Firstly,the thesis expounds the background and significance of this topic,and the development status of mobile robot at home and abroad and integrated navigation technology are discussed in this paper.Then it builds a multi-sensor integrated navigation platform based on SINS/GPS/OD,including inertial sensor module,encoder module and GPS module,and designs the hardware and software of some modules in detail.Among them,the hardware mainly includes attitude acquisition circuit,magnetic encoder module circuit and differential circuit,the software mainly includes the attitude algorithm of the strapdown inertial navigation system,collection of motor speed and the position of the robot and so on,finally it expounds the design method for application of the integrated navigation system,and gives the specific design scheme,which laid the foundation for the follow-up data acquisition and experimental demonstration.Secondly,the algorithm of quaternion and complementary filtering is described respectively for solving the attitude problem of two-wheel self-balancing mobile robot.The adaptive quaternion Kalman filter algorithm and the improved complementary filtering algorithm are proposed at the same time.The former reduces the non-gravitational acceleration disturbance and improves the attitude precision.The latter can improve the attitude convergence speed of the robot in the initial running phase and effectively guarantee the stability of the body.Thirdly,aiming at the nonlinear characteristics of the system,the error model of inertial system and odometer is established,and the Kalman filter integrated navigation method based on SINS/OD is simulated and compared with traditional dead reckoning method,and analysed the superiority of the former.Then,an adaptive federated filter based on SINS/GPS/OD is designed,and introduces its effect on eliminating the dynamic model error,it also can improve the localization accuracy of the robot under the model error surroundings.Finally,the thesis tests and demonstrates the function modules and overall functions of integrated navigation system in the two-wheel self-balanced mobile robot.Then,it tests the effect of the magnetic encoder for the forward and reverse rotation of the motor,and the system is combined with the inertial navigation module,also the performance of the integrated navigation system is tested by using the GPS module embedded in the loose combination method.Finally,the output data of the inertial sensor is collected to verify the effectiveness of the adaptive quaternion Kalman filter and the improved complementary filtering algorithm.