旋翼飞行机器人动平台自主起降方法研究

CORC > 沈阳自动化研究所 > 中国科学院沈阳自动化研究所 > 机器人学研究室

题名	旋翼飞行机器人动平台自主起降方法研究
作者	吴冲
学位类别	博士
答辩日期	2015-05-26
授予单位	中国科学院沈阳自动化研究所
授予地点	中国科学院沈阳自动化研究所
导师	韩建达
关键词	旋翼飞行机器人卡尔曼滤波扩展集员滤波线性规划动平台
其他题名	Researches on Autonomous Takeoff and Landing of an Unmanned Helicopter On a Moving Platform
学位专业	模式识别与智能系统
中文摘要	直升机所特有的垂直起降、定点悬停、低速低空飞行等能力，使得直升机作为现有海洋舰船装备中重要的组成部分，在海上侦察、海事救护、反潜反舰、两栖突击、空中预警等海事应用中具有不可替代的作用，极大地拓展了舰船在海上的应用能力。但是舰载直升机的起降涉及到飞行驾驶员与舰面指挥人员的紧密配合，需要通过舰面指挥人员的引导以完成安全起降，舰载直升机起降是危险系数最大的环节，也是事故率最高的环节。旋翼飞行机器人作为有人驾驶直升机的一种辅助替代方案，可以完成大部分目前有人驾驶直升机的作业任务。相比于有人驾驶直升机，旋翼飞行机器人的建造、使用、维护成本可以大幅下降，而且在应用过程中可以避免驾驶员人员伤亡，在未来海事应用中具有重要的作用。旋翼飞行机器人的舰载应用与常规应用最大的差异在于舰载自主起降，舰载自主起降的特点体现在三个方面：首先是舰面起降平台空间狭小，对旋翼飞行机器人的状态估计要求精度更高、估计结果更稳定，而且对舰船运动需要有一定的预测能力以指导路径规划的开展；其次起降平台处于运动状态，要求旋翼飞行机器人在动力学约束下安全跟踪起降平台的同时避免与舰船上其他建筑物相碰撞；另外起降平面处于实时摇摆状态，起降时机的选择以及对应的稳定控制直接影响到起降的安全。本论文将从以上三个方面着手，以舰船甲板平台为典型动平台，研究分析制约旋翼飞行机器人动平台自主起降的关键问题，建立旋翼飞行机器人动平台自主起降方法体系，鉴于舰载实验成本高风险大，还将通过搭建地面运动平台模拟舰船运动来完成对理论方法的飞行实验验证，最终实现旋翼飞行机器人在动平台上的自主起降。本文的主要研究内容如下：（1）旋翼飞行机器人与动平台状态估计方法研究针对旋翼飞行机器人动平台自主起降需求，从飞行器状态在线估计、飞行器健康状态在线估计以及动平台运动状态预测三方面内容展开研究：首先针对现有飞行状态估计算法维数过高导致的计算复杂度大、稳定性差的特点，提出了一种基于卡尔曼滤波的降阶飞行状态估计方法以实现在线实时求解，设计并搭建了完整的导航系统并经过飞行测试对算法的有效性进行验证。其次针对影响飞行安全的飞行器健康状态在线监测困难这一问题，提出了基于扩展集员估计滤波的健康状态在线估计方法，以实现对传感器和执行器故障的实时诊断与估计。最后结合动平台运动特点，提出了基于时间序列分析的动平台运动状态预测方法，实现了短时期的动平台运动状态预测。状态估计方法的建立为后续路径规划与飞行控制的研究提供了必要的支撑条件。（2）基于动力学约束的自主降落路径规划方法研究为确保规划路径的可行性与安全性，针对动平台自主降落的特点，在线性规划的框架下，建立了旋翼飞行机器人的动力学约束模型以确保规划路径的可行性，同时建立了动平台追踪优化模型和动态避障约束模型以确保规划路径的安全性，基于建立的线性规划模型可以进行在线实时求解以引导旋翼飞行机器人在动平台的降落。（3）基于安全性能分析的自主起降流程切换控制机制研究旋翼飞行机器人的自主起降流程涉及到多个模态的切换与控制，基于对近地地效的分析和起落架与动平台摩擦力的分析，结合飞行器、飞机健康因子的状态估计以及动平台运动的预测，建立了旋翼飞行机器人自主起降切换控制机制，可以在现有飞行控制器结构的基础上，确保安全的自主起降。（4）旋翼飞行机器人动平台自主起降实验研究在上述理论研究的基础上，首先搭建了半物理仿真系统完成了对路径规划、自主起降飞行切换控制机制的仿真验证，而后搭建了地面运动平台以模拟舰船在不同海况下的运动，旋翼飞行机器人基于该运动平台进行了飞行实验验证。在飞行实验验证阶段分别完成了定点自主起降、定点摇摆平台自主降落、动平台跟踪飞行测试与动平台自主降落等科目，验证了理论方法的可行性。本文的研究工作对旋翼飞行机器人的动平台自主起降进行了有益的尝试，初步验证了动平台自主起降的可行性，有助于推动旋翼飞行机器人的实际舰载应用。
索取号	TP242/W81/2015
英文摘要	Due to the specified capacity of vertical takeoff and landing, fix point hovering, and low speed flight at low altitude, helicopter has become an indispensible equipment for various kinds of ships, it has been widely used in sea reconnaissance, maritime rescue, antisubmarine, antiship, amphibious operations, and aerial pre-warning. Shipboard takeoff and landing of a helicopter on a ship was based on the closely work between the pilot and the instructor on the vessel, instructor on ship was neccessary because pilot can not know comprehensively about the conditions on the ship, takeoff and landing was the most dangenrous stage for the shipboard operation of helicopters. As a substitution of shipboard helicopter, unmanned helicopter can finish most of the task that should be finished by piloted helicopter, and it is much cheaper and easier to be maintained without any casualty, unmanned helicopter could be a very important shipboard equipment in the future. Compared with the traditional unmanned helicopter, the main difference and chanllenge is the autonomous shipboard takeoff and landing, there are mainly three aspects should be concern: firstly, the takeoff and landing area is very limited on the vessels, the unmanned helicopter should be more awareness about the flight state and the ship state; secondly, ship deck is moving all the time, the unmanned helicopter should be able to follow the movement of the ship deck and avoid the abstacles around; thirdly, the swing of ship’s deck on the sea means that the touch point of unmanned hehlicopter and ship would be at any condition, how to control the unmanned helicopter to touch the deck at a relatively safe position and attitude is critical for the operation safety. This thesis will take the ship as a special moving platform, try to find out the limitations existed for the autonmous takeoff and landing of shipboard unmanned helicopter, build a research framework to tackle these limitations. The main work of this thesis can be summarized as follow: (1) State estimation for unmanned helicopter and moving platform State estimation was divided into three aspects: flight state estimation, healthy state estimation, and moving platform movement prediction. Firstly, a cascased extended kalman filter was proposed for the flight state estimation to avoid high dimentional matrix calculation existed in the traditional flight state estimation method, an integrated navigation hardwared and software system was established to verify the effctiveness of this method. Secondly, extended set-membership filter was introduced for online sensor and actuator fault estimation. Thirdly, based on the characteristics of the moving platform’s movement, a time series analysis method was proposed for the moving platform prediction. State estimation was the fundation of online path planning and flight control. (2) Dynamics contrained path planning for autonomous landing In order to gurantee the safety and feasibility of the planned path, unmanned helicopter’s dynamics constraints were transformed into linear contraint for the path planning. At the same time, target pursuit and obstacle avoidance rule and contraint were built respectively in the linear path planning framework. Based on the linear path planning model established, a feasible path can be generated to guide the unmanned helicopter autonomous landing. (3) Safety analysis based controller designed for autonomous take off and landing As the unmanned helicopter executes autonmous takeoff and landing, many stages and mode would be covered with different Dynamic Characteristics. Based on the analysis about the ground effect and the friction between helicopter and moving platform, a flight control mechanism was built up to guarantee the stability of shipboard autonomous takeoff and landing. (4) Flight test and validation for the unmanned helicopter shipboard autonomous landing In order to validate the effectiveness of the proposed method, a hardware in the loop simulation platform was established to verify the proposed path planning and control mechnism algorithms, besides that, a moving platform was established to simulate the movement of ship on the sea, and flight tests was conducted based the moving platform. Autonmous takeoff and landing on the fix platform, fix but swing platform, and moving platform wer all finished to verify the algorithms proposed in the thesis. The research in the thesis was an useful attempt of the shipboard autonmous takeoff and landing of unmanned helicopter, the feasibility of shipboard autonomous takeoff and landing has been verifed which will promote the real shipboard operation of unmanned helicopter.
语种	中文
产权排序	1
页码	127页
内容类型	学位论文
源URL	[http://ir.sia.ac.cn/handle/173321/16808]
专题	沈阳自动化研究所_机器人学研究室
推荐引用方式 GB/T 7714	吴冲. 旋翼飞行机器人动平台自主起降方法研究[D]. 中国科学院沈阳自动化研究所. 中国科学院沈阳自动化研究所. 2015.