【作者】 徐玉；

【导师】 李平； 韩波；

【作者基本信息】 浙江大学 ， 控制科学与工程， 2008， 博士

【摘要】 无人机可以在“枯燥任务领域、恶劣环境任务领域和危险任务领域”三个特定的环境领域发挥出重要作用。由于具有体积小、成本低、可垂直起降和定点悬停、机动灵活、隐蔽性好、无人员伤亡、战场生存能力强等优点,微小型无人直升机近年来已经成为国内外的研究热点。在微小型无人直升机的研究中,飞控平台的构建是实现自主飞行前急需解决的关键问题。本论文的主要工作是针对微小型无人直升机对载重、成本敏感的特点,系统而完整地给出了低成本飞控平台和姿态融合算法的原理设计、工程实现及实验验证,包括:飞控系统硬件选型与集成,传感器校正,基于GPS加速度补偿和滤波切换的低成本MARG传感器姿态融合算法的设计、实现及实验验证,速度位置估计算法的设计与实现,以及发动机转速定速器的设计与实现。论文结合低成本MARG传感器的特点,重点研究了无人直升机姿态融合算法的设计。论文的研究成果为微小型无人直升机的辨识建模与自主飞行实现提供了有力保障。论文的主要工作为:1.介绍了论文的研究背景与意义,概述了微小型无人直升机的国内外研究现状,然后对无人机姿态测量技术进行了综述,最后给出了论文的研究目标、内容与篇章结构。2.给出了飞控系统的硬件设计与实现方案,介绍其基本功能;然后介绍了关键传感器的误差模型,对传感器的误差源进行了分析;最后针对捷联传感器的误差特点,提出了一种基于递推最小二乘的椭球校正法,并给出了实验结果,该方法具有一定的创新性和实用性,非常适合低成本捷联加速度计、磁传感器的校正。3.针对基于重力场和地磁场矢量观测的低成本MARG传感器姿态测量系统易受载体运动加速度影响的特点,提出一种用GPS加速度补偿运动加速度以提高载体机动时的姿态测量精度的方法。首先利用三阶最佳差分器从GPS速度差分获得GPS加速度,并对其测量精度进行了实验验证;然后利用GPS加速度与加速度计输出一起构建了比力矢量观测方程;最后给出了基于修正罗德里格参数的姿态融合算法和基于四元数Kalman滤波器的姿态融合算法;前者利用迭代最小二乘姿态确定算法从比力矢量和地磁矢量观测方程得到参考姿态,该参考姿态与陀螺仪输出进行互补滤波融合,其中着重解决了观测滞后与修正罗德里格参数姿态表示方法的奇异问题;后者通过构建线性虚拟观测方程,将带观测滞后的非线性系统滤波问题转化成带时间相关噪声与状态相关噪声的线性系统滤波问题并给出了相应的Kalman滤波算法。4.针对基于重力场和地磁场矢量观测的低成本MARG传感器姿态测量系统易受载体运动加速度影响的特点,提出了一种基于滤波切换的姿态融合算法。利用两个互补滤波器分别估计重力场与地磁场,并分析了互补滤波器的性能;为减小运动加速度对重力场估计的影响,在载体作机动时将重力场滤波器切换到陀螺仪开环积分;分析表明,在一定的切换条件下,切换滤波系统的估计误差是有界的;将该算法和第三章两种算法进行了仿真和实验研究,结果表明三种算法估计的姿态均具有良好静态精度和动态性能,与低成本MARG传感器结合均能满足无人直升机自主飞行的要求。5.在获得姿态的基础上,采用了分级式GPS/SINS组合导航结构,给出了基于互补滤波器的速度位置估计算法;该算法估计无人直升机相对地面站的速度与位置,实现简单、计算量低,适用于短航程无人直升机应用。实验结果表明,该算法都能使分级式GPS/SINS组合导航系统具有较高的导航定位精度,可以满足无人直升机自主飞行的要求。6.给出了一种基于前馈-模糊自整定P1控制的无人直升机发动机转速定速器的设计与实现方案;并将此发动机转速定速器应用于OS MAX 70SZ、YS 91 ST和Rotax 582发动机的转速控制,实验结果表明,被控发动机在有负载干扰的情况下能很好地保持转速稳定,具有良好的抗干扰性能。该方法实际可行、易工程化,可用于其他无人直升机的发动机控制。7.总结了论文的研究成果,并对未来的研究工作做了展望。更多还原

【Abstract】 The unmanned aerial vehicle (UAV) has been playing an important role in "dull, dirty, or dangerous" missions where human intervention is considered difficult or dangerous. Due to the advantage of small size, low cost, vertical take off and landing ability, hovering ability, maneuverability, good concealment ability, no casualties and strong viability in the battlefield, the study of mini autonomous helicopter (MAH) has been one of the most interesting research fields in the world.The development of flight control platform is the linchpin before the realization of MAH’s autonomous flight. The main work of this dissertation is giving the systematic and complete theoretic design, realization and experimental validation of a flight control platform and attitude fusion algorithms according to the low cost and low weight requirement for MAH, including the hardware selection and integration of flight control system, sensor calibration, design, realization and experimental validation of two GPS acceleration compensation based low cost MARG sensors attitude fusion algorithms and a low time consuming MARG sensors attitude fusion algorithm based on switch technology, design and realization of cascade GPS/SINS integrated navigation system and engine speed controller for mini UAV. The key parts of this dissertation are the research on attitude fusion algorithms for low cost MARG sensors. We believe that the outcome of this dissertation can provide powerful ensurence for system idenfication and autonomous flight control of MAH. The main work of this dissertation is:1. The background and significance of this dissertation are presented and the current progress of MAH is introduced, then a brief introduction to attitude determination technologies of UAV is given, and finally the research purposes, contents and organization of this dissertation are shown.2. The hardware design and realization scheme of flight control system is presented, then the error models of key sensors are analized, finally a recursive least squares based ellipsoid hypothesis calibration algorithm suitable for low cost strapdown accelerometers and magnometers is presented.3. Because the translational acceleration may affect accelerometer output and result in performance reduction of accelerometers and magnetometers based low cost MARG sensors attitude determination system, the GPS-derived acceleration is introduced to improve the attitude measurement accuracy during maneuvers. Firstly, the translational acceleration is derived from GPS velocity through three order optimal differentiator and its accuracy is experimental validated. Then the GPS-derived acceleration and accelerometers’ output are used to form the specific force vector observation equation. Finally two GPS acceleration compensation based low cost MARG sensors attitude fusion algorithms are proposed. The first algorithm employs an iterative least square attitude determination algorithm to determine the MRPs attitude from the specfic force and magnetic field vector observations, then MPRs attitude is then fused with the measurements of rate gyros using complementary filter, the observation delay and singularity of MRPs are also considered. The other algorithm employs linear pseudo-measurement equation to transform the nonlinear system with delayed measurement to a linear system with state dependent noises and temporal correlated observation noises, and gives the corresponding linear kalman filter design.4. A low time consuming MARG sensors attitude fusion algorithm based on switch technology is presented. The algorithm is composed of two complementary filters that estimate gravity and magnetic field vectors, and the filter performance is analyzed. To reduce the undesirable effect of translational acceleration, the gravity field complementary filter may switch to open loop integration using gyros’ output during maneuver. Analysis indicates that the estimation error of the proposed switched system is bounded under certain switching condition. Simulations and experiments have been carried out to verify the performance of the proposed algorithms in Chapter 3 and Chapter 4, both simulation and experimental results show that the three algorithms are superior in steady state accuracy and dynamic performance, and can fulfil the autonomous flight requirements of the unmanned autonomous helicopter using low cost MARG sensors.5. The cascaed GPS/SINS integrated navigation structure is adopted on the basis of attitude estimates in Chapter 3 and Chapter 4, and a velocity and position estimate algorithm using complementary filter is given. The algorithm employs two complementary filters to estimate velocity and position of the MAH relative to ground station. This method is easy to realize, low time consuming and suitable for short voyage MAH. The experimental results show that the algorithm possesses good navigation accuracy and can fulfil the autonomous flight requirements of MAH.6. The design and implementation scheme of an engine speed controller that adopts the control algorithms of feed forward and fuzzy tuned PI is presented. The proposed engine speed controller is then applied to the engine speed control of OS MAX 70SZ, YS 91 ST and Rotax 582 engines, the experimental results have shown that the controlled engine possess excellent disturbance rejection peformance and can hold its speed constant well under load disturbance. This method is feasible and can be benfit for engine speed control of other MAHs.7. The main work of the dissertation is summarized and recommendations for future research are presented.更多还原