【作者】 王丹；

【导师】 张卫忠；

【作者基本信息】 北京理工大学 ， 航空宇航科学与技术， 2015， 硕士

【摘要】 四旋翼飞行器是典型的集强耦合、非线性、多变量等特性为一体的旋翼式无人机。目前，针对该飞行器的研究涵盖系统结构设计、模型建立、控制律设计和路径规划等。其中，为了实现飞行器在实际情况中规避危险以完成搜索或跟踪任务，并结合自身负载的局限性，能够事先规划好尽量短时间的飞行轨迹并进行准确地跟踪控制就显得尤为重要。由于四旋翼自身的特性以及飞行任务的复杂性，为了进一步提高算法地适用性、收敛速度以及可解性，本文针对四旋翼飞行器，在现有轨迹优化算法和跟踪控制算法的基础上进行适当改进，并通过大量的仿真实验加以验证分析。主要研究内容如下：1.建立四旋翼飞行器模型。在合理假设的基础上，建立双坐标系。依据飞行器特性分析，结合牛顿第二定理以及欧拉方程，推导系统相应的动力学方程。同时，为了后续轨迹优化及控制器设计工作的顺利进行，对数学模型进行了线性化和解耦处理。2.提出借助于辅助初值猜测策略（Auxiliary Initial Guess Strategy, AIGS）的MDMOC (Multi-phase Discrete Mechanics and Optimal Control)轨迹优化算法。通过传统DMOC方法、高斯伪谱法(Gauss Pseudo-spectral Method, GPM)和一种借助于AIGS结合分段轨迹优化策略的MDMOC方法的综合对比，得出：就四旋翼飞行器问题而言，GPM要优于DMOC法，借助于AIGS的MDMOC在优化效率和程序化方面优于GPM；同时，MDMOC在保证一定精确度的同时，更加适用于复杂飞行要求和飞行环境中使用。因此，采用MDMOC将作为下一部分生成控制器设计中最优参考输入轨迹的优化方法。3.设计飞行器自适应轨迹跟踪控制系统。四旋翼飞行器复杂的动力学特性和诸多的不确定性使得跟踪控制器的设计变的更加困难。因此，分别采用经典PD控制算法和一种质量未知情况下基于反步法的滑模变结构自适应控制算法设计轨迹跟踪控制器。仿真结果显示，两种方法均能实现较为稳定的飞行和轨迹跟踪。具体的，在系统达到稳定所需的上升时间，稳定时间以及鲁棒性方面，基于反步法的滑模变结构自适应算法具有较大的改善，更加适用。更多还原

【Abstract】 Quadrotor is a kind of typical rotor aircraft which model is strong coupling, nonlinear,multivariable and under-actuated. At present, the main research interests consist of systemstructure design, modeling, control and path planning, etc. In practice, in order to realizerisk aversion to search or tracking with the limitations of its load, it is particularly importantto preplan a minmum time trajectory and design the accurate tracking controller.Due to its features and the complexity of the mission, in this paper, we propose someappropriate improvements on the basis of the existing trajectory optimization algorithm andthe tracking control algorithm to improve the convergence speed, solvability andapplicability. Then, some simulation experiments are verified and analyzed. The mainresearch contents are as follows:In this paper, the optimal trajectory generation and tracking control of the quadrotorare mainly researched by a large amount of simulation experiments to verify the proposedmethods. The main contributions of the paper are as follows:1. The establishment of the quadrotor model. The reasonable hypothesis is introduced anda double coordinate system is set up. Via analyzing the characteristic of the quadrotor andaccording to the Newton’s law and Euler equation, we deduce the corresponding dynamicequations. At the same time, for the sake of the subsequent trajectory optimization andprogress of the controller design, the mathematical model is linearized and decoupled.2. The MDMOC algorithm and the AIGS is put forword. By the comprehensivecomparison of the traditional DMOC, GPM and the proposed MDMOC, we finally drawthat the MDMOC with AIGS in optimization efficiency and procedural aspects are betterthan that of GPM. What’s more, MDMOC is more suitable under complex requirementsand flight environment with certain accuracy. Therefore, MDMOC will serve as the methodof generating optimal reference trajectory in the following controller design.3. Adaptive tracking controller design of the quadrotor. Because of the complicateddynamic characteristics and many uncertainty makes the tracking controller design of thequadrotor becomes more difficult. This paper is based on the classic PD control algorithmand the backstepping method with sliding mode variable structure adaptive controlalgorithm under unknown quality to design tracking controller respectively. The simulationresults show that both control algorithms can achieve a stable tracking control. Specifically, as for the robustness, the rise time and the stability time of the system reach a steady state,the backstepping method with sliding mode variable structure adaptive control algorithmhas great advantage.更多还原