| The Multi-rotor UAV(unmanned aerial vehicle),designed with new structures,has received more and more attention for its simple structure,easy operation and cost-effective production.Playing important roles in military surveillance,forensic photography,industrial inspection,pesticide spraying and monitoring,commercial aerial shooting and selfie,it has become hot research topics and been applicated in multiple fields.Accurate sensor measurements and information fusion are indispensable for steady flight of multi-rotor UAV.Hence,this paper puts forward and solves following technical problems in consideration that attitude,velocity and position of the mirco-rotor UAV should be exactly estimated under complex conditions near the ground.The first problem is how to improve accuracy and reliability of vehicle-borne navigation systems,and the second problem is how to accurately and reliably estimate the attitude and position of the mirco-rotor UAV under complicated conditions.In view of these technical problems,specific research content and innovations of this paper are summed up as follows.For improving measurement accuracy and reliability of the MEMS gyroscope in the vehicle-borne navigation systems,it is proposed that G-sensitivity error calibration method of the MEMS gyroscope,to compensate dynamic errors of the gyroscope resulting from acceleration.The calibration scheme based on(8+4N(N?1))symmetric turntable positions is designed,in order to effectively decouple the impacts of random noise and bias of the gyroscope,and the earth's rotation rate upon G-value sensitivity error calibration accuracy.Thereby,G-sensitivity error of the MEMS gyroscope is calibrated and compensated,so as to improve the dynamic performance of the gyroscope.Based on that,a heterogeneous orthogonal redundancy collocation method is proposed by additionally equipping industrial MEMS gyroscopes with a type of cheap wide-range gyroscopes in consideration that cheap,accurate and wide-range measurements of MEMS IMU are necessary for multi-copter unmanned vehicles.Besides,an adaptive fusion algorithm is designed for redundancy collocation,so that measurement properties and ranges of the MEMS gyroscopes are improved and expanded without increasing costs and volumes of micro inertial measurement units.In addition,to improve dynamic measurement properties of MEMS magnetometer in vehicle-borne navigation systems,an online calibration method is proposed for the magnetometer,in view that the traditional offline calibration method is difficult to operate in some large multi-rotor unmanned aerial vehicles and the error parameters calibrated offline are impossible to adapt to changes of the magnetometer error model during the flight.In the online calibration method,the attitude of UAV is used for dividing different ellipsoid quadrant,so as to select the magnetometer output for fitting the the error parameters.And such error parameters are fitted by the minimal eigenvalue method.Thus,error compensation accuracy and dynamic adaptability of MEMS magnetometer are effectively improved.In consideration that the accuracy of the multi-rotor UAV attitude fusion algorithm is impacted by the acceleration of UAV,an attitude fusion algorithm is put forward for online adaptively estimating acceleration,and an adaptive algorithm is designed for regulating the decay coefficient of acceleration models.As a result,horizontal attitude fusion accuracy and reliability are effectively improved by isolating the interruption of acceleration.Subsequently,a heading correction method based on a second-order damping circuit is explored,and the estimation accuracy of the heading is thus improved.The proposed attitude fusion algorithm would be effective for providing stable reliable backups about attitudes,so as to maintain stable attitude control of the multi-rotor UAV,if the position and attitude fusion algorithm could not meet requirements for navigation control of UAV.To guarantee reliable and stable flight of the multi-rotor UAV,in addition to researching the independent attitude fustion algorithm,it is still necessary to further investigate stable and reliable attitude and position fusion algorithm.In view that there are some problems with attitude and position fusion algorithm such as strong nonlinearity,unstable measurements and time variances of measured noises when the multi-rotor UAV flies near the ground,an optimized SHA-UKF adaptive fusion algorithm is put forward based on linear and nonlinear measurements update.Besides,traditional evidence for judging filter divergence is improved based on self-stable evidences and monitoring of abnormal measurements.Thus,accuracy and reliability of UAV attitude and position estimation are effectively improved.At last,a hardware-in-the-loop computing platform is designed for fusion algorithms on the premise that the multi-rotor UAV are equipped with EMPU,for the purpose of making performance validation of attitude and position fusion algorithms more authentic and practical.In addition,fusion algorithms are run again by the hardware reconstruct according to collected data about flight experiments and results are comparatively analyzed.It has been fully verified that the fusion algorithms proposed in this paper are effective for improving accuracy and stability of attitude and position estimation for the multi-rotor UAV with high value for engineering practices.Methods investigated in this paper for calibrating errors of MEMS sensors,improving their measurement performances and fusing attitude and postion information cannot only effectively improve measurement accuracy and reliability of the vehicle-borne navigation systems,but also can realize accurate and reliable measurements of the attitude and position under complicated conditions near the ground.These methods provide important technical references for improving performances of navigation systems in the multi-rotor UAV. |
PDF Full Text Request