节点文献
轻型飞行模拟器运动平台先进控制技术研究
Advanced Control Technology Research of Light Flight Simulator Motion Platform
【作者】 吴东苏;
【导师】 顾宏斌;
【作者基本信息】 南京航空航天大学 , 载运工具运用工程, 2007, 博士
【摘要】 近年来,随着民航业的不断发展,飞行培训的需求呈现日益强劲的态势,面对传统飞行模拟器高昂的采购价格和使用成本,航空界一直在寻求一种功能与传统飞行模拟机相当,但购置和使用成本低廉的经济型飞行模拟器。为此,我们提出了“轻型飞行模拟器”的概念,这是一种轻巧紧凑、以软件代替硬件的低成本模拟器,现在已经完成了原理样机的设计和制造。高性能的六自由度运动平台是这种轻型飞行模拟器系统中的一个重要组成部分,它的设计目标是能在普通层高的建筑物中进行布置,结构简单,重量轻;利用电动执行器实现较好的运动性能和加速度能力,能适应人员和装备的变化;能对运动感觉进行高性能的模拟;同时配备完善的软件限制保护措施,使之能进行自我保护,并尽可能的不影响仿真效果或产生仿真中断。本文针对运动平台的目标要求设计完成了一个完整的轻型飞行模拟器运动平台控制系统,对并联平台的建模、模型参数辨识、平台控制、动感模拟算法的设计优化和运动保护系统的设计等一系列问题展开了研究。本文的第一部分内容主要集中于运动平台的建模、辨识与控制。首先对三种动力学建模方法进行了比较研究,对比了它们的计算复杂度和应用场合等并研究了它们之间的等价关系。然后提出了使用UKF方法对并联运动平台进行参数辨识的方法,可以直接对非线性模型进行处理,而且精度很高。针对并联运动平台存在不确定性参数和外部扰动的情况,本文又提出了一系列自适应控制算法保证运动平台的控制性能。首先设计了自适应滑模控制器将平台的不确定性分为定常和时变的进行分别处理,然后进一步针对滑模控制器的高增益项会引起抖振的问题,又设计了基于模糊干扰观测器的自适应模糊控制器。接下来本文在前面所设计的自适应模糊控制器的基础上将执行器动力学的影响考虑进控制算法中,设计了运动平台的自适应模糊反步控制算法,可以保证整个系统的渐进稳定,同时保持良好的跟踪性能和抗干扰性能。最后针对电动飞行模拟器并联运动平台电动执行器有限的驱动力,提出了一种针对执行器饱和的轨迹跟踪算法。这种方法结合使用了实时轨迹修正、带饱和补偿的模糊干扰观测器和分支速度观测器,使得存在外部干扰和无分支速度测量情况下,平台控制器始终工作在非饱和状态,保证了控制器的稳定并保持了良好的跟踪特性。本文的第二部分内容设计和优化了轻型飞行模拟器的动感模拟算法。针对轻型飞行模拟器的特点选用了经典洗出滤波器,然后本文利用运动平台可达工作空间和所需工作空间需相匹配的原理,优化了动感模拟算法的参数,以保证能充分利用有限的并联平台运动空间达到逼真的动感模拟效果。本文的第三个部分设计了轻型飞行模拟器的运动保护系统。在分支关节空间设计了杆长限制算法,并针对任务空间运动在缓冲过程中容易发生错误运动这一问题对算法进行了改进。同时对于分支杆与平台座舱之间的干涉和关节摆角干涉这类问题,又提出了任务空间的虚拟弹簧干涉缓冲算法,假设平台构件表面都覆盖有虚拟弹簧,一旦构件相互接近到一定程度,会产生虚拟弹簧力对接近运动进行阻碍。本文的最后一部分针对轻型飞行模拟器设计了完整的性能测试的环境和流程,给出了性能指标以及最终的测试结果。性能测试结果表明运动平台具有良好的静态和动态运动性能。本文控制器设计都经过了严格的证明,并通过大量的计算机仿真试验和数据验证了控制律的正确性和有效性。所述的内容完整的解决了飞行模拟器运动平台控制的一系列关键问题,也为今后平台性能的进一步改进打下了坚实的基础。文中所设计的运动平台控制系统已应用在“轻舟一号”轻型飞行模拟器原型系统中,实际运行良好。
【Abstract】 Nowadays, with the rapid development of civil aviation industry, there are increasingly strong demands of pilot training. But due to high purchase price and usage costs of tranditional flight simulator, a low-cost substitution is expected to appear that can cover the basic functions of the tranditional one. In order to meet the urgent needs of civil aviation industry, we propose a concept of“Light Flight Simulator”. The most exciting feature of this economic simulator is its low weight and massive usage of software modules to fulfill the functions that tranditional hardware must do. Now we finish design and manufacture of the prototype simulator. One of the most important parts of this simulator is a high performance six-dof motion platform. Its design goals are simple structure, low weight and capable of dispose within common office building. At the same time, the simulator must have good dynamic ability, supply satisfying motion feeling for pilots using electrical actuators and be capable of adapting variation of equips and pilots. Moreover, the simulator must be equipped with full software limiting measures, which can protect the hardware from accidents. This dissertation designs an integrated control system framework for light flight simulator motion platform, which covers a series of research subjects including modeling, identification, control of motion platform, washout algorithm design and optimization, motion protection system design and etc.The first part of this dissertation covers modeling, identification and control of motion platform. Firstly, three main dynamic modeling methods are compared in its computation complexity, application area and equivalence. Secondly, UKF method is proposed to identify parameters of parallel motion platform. Comparing to other identification methods, UKF method does not need transforming dynamic model to linear parameter form and have higher identification precision. Due to uncertain parameters of parallel motion platform and external disturbance, this dissertation proposes a series of adaptive control schemes to ensure the control performance of motion platform. The first controller proposed is a nonlinear adaptive sliding control scheme in task space based on dynamic model. This method divides system uncertainty into two parts: constant uncertain parameters and time-varying uncertain parameters. The control algorithm uses nonlinear adaptive controller to identify constant uncertain parameters, meanwhile, nonlinear sliding controller is used to compensate the effects of time-varying uncertain parameters and external disturbance. Due to inclusion of high-gain part in the adaptive sliding controller, it may excite unmodelled dynamics and cause chattering. So this dissertation proposed another novel adaptive fuzzy algorithm based on fuzzy disturbance observer to solve this problem. Next we step afurther and incorporate actuators’dynamics into controller design process to get an adaptive backstepping controller. This controller can ensure global stable of the whole system, good tracking ability and disturbance rejection ability. Finally due to limit actuator ability of electrical driven flight simulator parallel motion platform, this dissertation proposes a novel trajectory-tracking scheme considering actuator saturation. This scheme incorporates realtime trajectory shaping, fuzzy disturbance observer and joint space velocity observer with saturation compensation. This kind of controller structure can ensure the motion platform always operating in non-saturation region, stable and have good tracking performance under external disturbance and no-velocity measurement situation.The second part of this dissertation designs and optimizes washout algorithm for light flight simulator. Classical washout algorithm is Chosen and parameters are optimized according to the theory of matching the available workspace and demand workspace. This method ensures effective use of limit motion platform workspace to get good motion simulation effect.The third part of this dissertation designs a novel motion protection system for light flight simulator which is used to handle leg length limit and avoid interference between legs, cabin and joints. A leg length limit algorithm is developed to prevent actuators from exceeding position, velocity and acceleration limits and reduce false motion in limit region. And an interference reduction algorithm in task space based on virtual spring is proposed to handle other constraint such as interference between legs, cabin and joints. This algorithm covers virtual spring on the surface of all structure parts. As soon as the distance between structure parts is within the limit region, virtual spring force proportional to distance will push those parts away.The last part of this dissertation designs complete performance testing environment and process for light flight simulator. The test results show that motion platform has good static and dynamic motion performance.All the controllers designed by this dissertation are proven strictly and lots of simulations and experiment datas verify their validity and efficiency. The contents of this dissertation solve a series of key problems related to effectively control of flight simulator motion platform and lay a solid base to further improve the performance of motion platform. The controller software module, washout algorithm software module and motion protection software module in this dissertation have been applied in“light boat No. 1”light flight simulator prototype we designed, and have good performance in daily use.
【Key words】 flight simulation; motion platform; Stewart platform; UKF; adaptive sliding control; adaptive fuzzy control; backstepping control; actuator saturation; washout algorithm; motion protection;