【作者】 杜建福；

【导师】 吕恬生；

【作者基本信息】 上海交通大学 ， 机械电子工程， 2008， 博士

【摘要】 直升机可以定点悬停、垂直起降,具有其它飞行器所不具有的突出优点,广泛应用在航拍、侦察等各个领域。小型无人直升机体积小、质量轻、成本低、结构简单,加载自主飞行控制系统以后还可以实现超视距飞行,能够完成许多常规飞行器所不能完成的任务。研究开发无人直升机自主飞行控制系统具有非常重要的现实意义。然而由于无人直升机本身是一个典型的欠驱动、强耦合、时变、有操纵滞后的非线性系统,开发无人直升机自主飞行控制系统是一个颇具挑战性的任务。针对当前小型无人直升机飞行控制研究中存在的问题,本文以一款单旋翼带尾桨式小型航模直升机为研究对象,设计开发了一套基于显式预测控制(EMPC)算法的小型无人直升机自主飞行控制系统,并通过飞行试验进行了实飞验证。论文的主要工作如下:1)小型无人直升机建模和大型直升机相比,小型无人直升机没有挥舞铰,其主旋翼具有更强的刚性和更高的转速,且主旋翼惯性效应非常明显,因此在建模中不仅将机体视作刚体,还将主旋翼也视作刚体,采用双刚体建模。论文以便于旋转运动学线性化的Yaw-Pitch-Roll(YPR,航向-俯仰-横滚角)为欧拉角,采用便于计算机求解的Kane方法详细推导了小型无人直升机的运动学方程和动力学方程,并对模型进行了分析和简化。2)飞行控制器设计针对常规无人直升机控制器不能有效处理执行器约束、输出约束以及操纵滞后等问题,本文采用预测控制算法实现无人直升机的飞行控制。但传统的预测控制需要在线优化,因此需要较大的计算量。为此论文引入显式预测控制算法。显式预测控制算法将在线优化问题离线化,在线计算时只需要查找相应的控制区域,通过简单的计算就可以得到控制量,因此非常适合嵌入式系统应用。小型无人直升机飞行控制器分为内环和外环,内环完成姿态控制,外环实现位置、速度控制。论文采用EMPC算法完成了姿态控制器、速度控制器和位置控制器的设计。3)鲁棒稳定性研究直升机在飞行中常会受到各种干扰的影响,因此控制器的鲁棒性就成为衡量控制器优良的重要指标。为考察EMPC的鲁棒稳定性,论文研究了飞行控制器在控制信号有干扰、机体变质量、外界风干扰等情况下的鲁棒稳定性,并和PID控制器进行了对比分析。4)路径跟踪能力验证为了验证飞行控制器的路径跟踪能力,文中通过仿真实验验证了飞行控制器跟踪矩形路径和“8”字型路径的能力。仿真实验表明,相比PID控制EMPC控制器具有较好的路径跟踪能力。5)实飞试验为验证飞行控制器的实际飞行性能,论文首先对实飞试验中需要知道的一些具体物理参数进行了辨识,然后在飞行实验室内进行了实飞试验。实飞试验表明,无人直升机姿态稳定,其俯仰角的变化范围在-2.9°~2.9°以内,横滚角的变化范围在-2°~2°以内,航向角变化不超过-2.9°~4.6°;无人直升机位置控制精度良好,其纵、横向通道位置精度达到0.15m,高度控制达到0.05m,完全可以满足工程应用需要。6)转向控制中的去耦问题针对直升机在转向过程中影响平面运动控制精度的问题,论文采用补偿算法对控制器进行了改进,并利用仿真和实飞试验进行了验证。仿真和实飞试验都表明,改进的预测控制算法能够较好的抑制这种耦合,具有较好的控制效果。围绕小型无人直升机控制器设计,本文从以上六个方面进行了研究,其主要贡献包括:论文提出将显式预测控制应用于6自由度小型无人直升机飞行控制;以YPR为欧拉角,采用Kane方法创建了小型无人直升机数学模型;基于所建模型和显式预测控制算法设计了小型无人直升机姿态控制器、速度控制器和位置控制器;另外,对于转向运动和平面运动之间的耦合论文还提出采用补偿算法进行改进。论文通过仿真和实飞试验验证了算法的有效性,本文所提方法在理论上和应用上都具有一定的借鉴作用和参考价值。更多还原

【Abstract】 Helicopters have superior flight characteristics that other aerocrafts haven’t, such as vertical take-off, landing and hovering, and are widely used in aviation photography, reconnaisance and other application fields. Small-scale unmanned helicopters have small size, light quality, low cost, simple structure and can fly beyond visual range and finish tasks that traditional aerospace can’t when furnished with autonomous flight control system. Thereby developing an autonomously flight control system is very significative and practical. However, helicopter models are under-actuated, strongly coupled, time-varying, multi-variate, time delay nonlinear, therefore, the design of an autonomous flight control system is a challenging task.In order to solve the problems encountered on current controller design, the author develops an autonomous flight control system based on explicit model predictive control (EMPC) for a single main rotor small-scale model helicopter with tail rotor. The contents of the paper will be discussed are listed as follows.1) Modeling of the small-scale unmanned helicopterCompared with large-scale helicopters, small scale unmanned helicopters haven’t flapping hinges, but have stiffer and higher speed main rotors. The inertial effect of the main rotor is very significant and the main rotor should be modeled as rigid body besides the fuselage. Kinematical and dynamical equations are deduced by Kane method which can be realized by computer easily. In order to linearize the rotational kinematical equation more easily, Yaw-Pitch-Roll (YPR) angles are chosen as Euler angles during modeling.The author also analyzes and simplifies the model.2) Design of the flight controllerConventional controllers can’t deal well with actuator constraints, output constraints and time delay, therefore the thesis proposes model predictive controller to control the unmanned helicopter. But ordinary model predictive controller needs online optimization and this leads to large computation complexity. Explicit model predictive control converts the online computation to offline and needs less computation complexity. It only looks up the control table and calculates the corresponding control input, so it is fit for embedded application. The flight controller consists of inner loop, which control the attitude, and outer loop, which control the position and velocity. The attitude controller, velocity controller and position controller are designed by the EMPC algorithms.3) Study on the robust stabilityHelicopters are usually disturbed during flight, so the robust stability of the flight controller is very important. In order to know the robust stability of EMPC, the controller undergoes control input noise, wind and the change of the mass of the helicopter. PID controller is also given for comparison.4) Validation of path tracking abilityIn order to validate the path tracking ability of the controller, the thesis simulate the controller to track rectangular path and "8" path. The simulation results show that explicit model predictive controller has better performance than PID controller.5) Real flight experimentsIn order to validate the performance of the flight controller in real environment, some real flight experiments are done in a flight lab. The results show that the helicopter can fly stably. The error of pitch angle doesn’t exceed -2.9°~2.9°, roll angle not exceed -2.0°~2.0°and yaw angle not exceed -2.9°~4.6°. The position precision is fine. The longitudinal and lateral error is 0.15 m and the vertical error is only 0.05 m. This is enough for application.6) Decoupling of steering controlA compensation algorithm is presented in this paper which is used to improve the precision of position control during steering. Simulations and real flight experiments validate its performance. The results show that the improved MPC algorithm works well and can restrain the coupling.The main contributions of the dissertation are as follows: EMPC is presented for the control of the small-scale unmanned helicopter; the mathematical model is deduced using Euler YPR by Kane method; the attitude controller, velocity controller, and position controller are designed based on the mathematical model and EMPC; in addition, the coupling between the steering motion and plane motion is dealt with compensations. The controller is validated by simulations and real flight experiments and the results show that the controller has good performance. The method given in the dissertation is worth to be used for reference in theory and application.更多还原