基于多传感器的多旋翼无人机导航解算方法研究

发布时间：2018-08-03 09:49

【摘要】：小型多旋翼无人机由于其便携、灵活、功能性强的特点,近些年受到了很大的关注。无人机的姿态控制系统需要得到准确的姿态信息作为反馈以便进行控制,同时要实现航迹跟踪、定点悬停等功能也需要准确的位置及速度信息。但无人机上空间有限,所搭载的导航传感器精度较低,测量噪声大,单靠某一个传感器的测量是无法得到准确的导航信息的。本文的研究重点就是通过设计导航解算算法,将多个传感器的测量信息进行融合,抑制单个传感器的噪声,从而获得较为准确的姿态、位置、速度信息,因此本文针对小型多旋翼无人机的导航信息解算的研究具有重要的研究价值和应用前景。本文首先对导航相关原理知识以及机上常用导航传感器的特性进行了介绍,包括常用导航坐标系、姿态角的表示方法、陀螺仪及加速度计的测量原理和噪声特性、磁强计及加速度计的姿态测量原理。然后,针对无人机在飞行过程中电机转动产生的高频振动会使陀螺仪、加速度计的测量值产生较大偏差的问题,研究并设计了数字低通滤波器,以便过滤掉传感器输出信号中机体振动所带来的高频噪声。本文通过分析机体高频振动对加速度计和陀螺仪的输出信号产生的影响,得到噪声的频率特性,并据此分别使用hamming窗函数法以及yulewalk函数法设计了FIR、IIR数字低通滤波器,并通过滤波效果以及滤波实时性的考虑,选择了可以有效抑制机体高频噪声幅值的数字低通滤波器。接着,针对如何准确地获取无人机的姿态信息的问题,设计了四元数形式的二阶互补滤波姿态解算算法并且对算法的解算效果进行了验证。该算法根据陀螺仪、加速度计、磁强计频率特性互补的原理,对三个传感器的输出进行数据融合,将陀螺仪作为主要测量元件,利用加速度计和磁强计对其进行修正,从而得到准确的姿态解算结果。同时,在Matlab中对算法程序进行了编译,搭建了仿真系统,分别使用仿真数据以及实验飞行数据在Matlab中对算法进行了验证。最后,本文针对如何准确获取无人机的位置、速度信息的问题,设计了基于扩展卡尔曼滤波器的位置、速度信息解算算法并且对算法进行了验证。该算法利用加速度计、气压计、GPS对机体飞行过程中的位置、速度进行测量,并且通过扩展卡尔曼滤波算法对三者的信息进行融合。本算法基于连续-离散时间系统,引入加速度计的偏移量作为状态向量,通过利用四阶龙格库塔法避免了雅克比矩阵的计算,提高了解算速度,最后利用仿真和实验数据对算法进行了验证。
[Abstract]:Small multi-rotor UAV has attracted much attention in recent years because of its portable, flexible and functional characteristics. The attitude control system of UAV needs accurate attitude information as feedback to control. At the same time, it also needs accurate position and velocity information to realize track tracking, fixed point hovering and other functions. However, the space of UAV is limited, the accuracy of navigation sensor is low and the measurement noise is large. It is impossible to get accurate navigation information by a single sensor alone. The focus of this paper is to design a navigation algorithm to fuse the measurement information of multiple sensors to suppress the noise of a single sensor, so as to obtain more accurate information of attitude, position and velocity. Therefore, the study of navigation information calculation for small multi-rotor UAV in this paper has important research value and application prospect. This paper first introduces the related principles of navigation and the characteristics of common navigation sensors on board, including the common navigation coordinate system, the expression method of attitude angle, the measurement principle and noise characteristics of gyroscopes and accelerometers. The principle of attitude measurement for magnetometers and accelerometers. Then, aiming at the problem that the high frequency vibration caused by the motor rotation of UAV during flight will make the measurement value of gyroscope and accelerometer deviate greatly, the digital low-pass filter is studied and designed. In order to filter out the high frequency noise caused by the vibration of the body in the output signal of the sensor. By analyzing the influence of high frequency vibration on the output signals of accelerometers and gyroscopes, the frequency characteristics of noise are obtained in this paper. Based on this, the Fir IIR digital low-pass filters are designed by using hamming window function method and yulewalk function method, respectively. Considering the effect of filtering and the real time of filtering, a digital low-pass filter which can effectively suppress the amplitude of high frequency noise is selected. Then, aiming at the problem of how to obtain the attitude information of UAV accurately, a quaternion second-order complementary filtering attitude algorithm is designed and the effect of the algorithm is verified. According to the principle of complementary frequency characteristics of gyroscopes, accelerometers and magnetometers, the algorithm fuses the output data of the three sensors, takes gyroscope as the main measuring element, and modifies it by using accelerometer and magnetometer. Thus the accurate attitude solution is obtained. At the same time, the algorithm program is compiled in Matlab, and the simulation system is built. The algorithm is verified in Matlab using the simulation data and the experimental flight data. Finally, aiming at the problem of how to accurately obtain the position and velocity information of UAV, the location and velocity information algorithm based on extended Kalman filter is designed and validated. The algorithm uses accelerometers, barometers and GPS to measure the position and velocity of the airframe during flight, and uses extended Kalman filter algorithm to fuse the three information. In this algorithm, based on the continuous discrete time system, the offset of accelerometer is introduced as the state vector. By using the fourth order Runge-Kutta method to avoid the calculation of the Jacobian matrix, the computational speed of the solution is improved. Finally, the algorithm is verified by simulation and experimental data.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V249.3;TP212

【相似文献】