QQ客服 官方微博 反馈留言

题名	无人机容错控制与重规划方法研究
作者	齐欣
学位类别	博士
答辩日期	2017-05-27
授予单位	中国科学院沈阳自动化研究所
授予地点	沈阳
导师	韩建达
关键词	容错控制 执行器故障与失效 执行器约束 规划与重规划 无人机
其他题名	The Research on Fault-tolerant Control and Re-planning for Unmanned Aerial Vehicles
学位专业	模式识别与智能系统
中文摘要	本文将以提升无人机的可靠性和智能化为目标，从理论角度对执行器故障后受约束被控对象的性能，及被控对象自身动力学约束与外部环境约束共同影响下的规划与重规划问题开展研究。本文主要研究工作如下： （1）受执行器约束系统发生执行器故障后的自治愈控制方法 执行器约束的存在限制了被控系统的性能，当发生执行器故障后，系统的可用剩余性能将进一步下降，增加了维持被控系统稳定性的难度。为此，将对现有容错控制方法进行扩展，提出自治愈控制方法框架，基于不变集思想，采用吸引域及可达集对故障后受约束系统的性能进行描述，从而判断执行器故障的可补偿性、故障后系统的稳定性及跟踪初始外部参考输入的能力。通过综合上述因素，设计容错控制器并规划新的最优外部参考输入，最大限度保证故障后被控系统的稳定性与安全性； （2）受执行器约束系统发生执行器卡死失效后的容错控制方法 执行器卡死失效作为执行器问题中较为严重的一种，其不仅减少了被控系统可用控制输入通道的数量，同时引入了不可控恒值输入，造成系统剩余可用性能既要用来补偿失效执行器又要用来抵抗恒值输入。因此，将在自治愈控制方法框架的基础上，提出针对执行器卡死失效的容错方法：基于补偿器的方法和基于执行器卡死与约束归一化描述的方法。其中，后者通过将执行器卡死失效放松并描述为执行器约束，构造了不包含卡死失效的新系统，提升了容错控制器设计的灵活性； （3）受自身动力学和外部环境约束的规划与重规划方法 无人机执行任务过程中，面临诸如障碍物、其他无人机等外部环境约束，需要通过规划方法计算同时满足外部环境约束及其自身动力学约束的轨迹才能确保无人机的安全性。针对这一问题，将首先基于自治愈控制思想，对现有的基于运动学约束的轨迹规划方法进行改进；然后，利用不变集既包含被控系统动力学模型信息又具有一定的几何形状这两个特征，将其作为连接被控对象自身动力学约束和外部环境约束的桥梁，提出基于不变集的规划思想，从理论上确保了被控对象运行过程中同时满足上述两种约束。在此基础上，对执行器发生故障后的重规划问题进行讨论； （4）室内试验平台搭建及面向四旋翼无人机旋翼失效的试验研究 基于运动捕捉系统搭建室内试验平台，以提升试验效率。基于上述理论思想，分析四旋翼无人机某一旋翼失效后的系统性能，并基于室内试验平台完成四旋翼无人机某一旋翼失效前后过渡模态飞行，及旋翼失效下的三维位置跟踪飞行。
英文摘要	As Unmanned Aerial Vehicles (UAVs) deploying in practical applications more widely, the requirements of improving their reliability and intelligence are increasing. However, all kinds of faults/failures is the main reason for influencing reliability and all kinds of constraints is an important factor limiting intelligence. In this paper, with the hope to improve the reliability and intelligence of UAVs, the problem of performance analysis of post-fault system with constraints and planning/re-planning under both dynamic constraints of the system itself and external environment constraints will be analyzed in theory. The main research results are as follows. 1) Self-healing control framework against actuator faults and constraints The system performance is limited by actuator constraints. Especially after actuator faults/failures, the remainder performance will be decrease which will increase the difficulty of maintaining system stability. Thus, this paper will extend fault-tolerant control approach to propose self-healing control framework. Based on invariant-set theory, self-healing uses attraction region and reachability region to describe the performance of post-fault system and these regions are used for evaluating the compensability of actuator faults/failures, the stability of the post-fault system, and the ability to track the initial external reference input. Then fault-tolerant controller and new optimal reference input will be designed according to the evaluation results to guarantee the greatest degree of stability and safety of the post-fault system; 2) Fault-tolerant control approaches against actuator stuck failure and constraints Actuator stuck failure is one of the most harmful actuator malfunctions. It not only decreases available control input channels but also brings into a constant disturbance. The remainder performance of post-fault system has to be used for both compensating the lost control input and rejecting disturbance. Thus, based on self-healing control, two approaches against stuck failure are proposed: compensator-based method and stuck-constraints-unified-description-based method. The latter one transforms stuck failures into actuator constraints to construct a new system without stuck failure so that it can increase flexibility in fault-tolerant controller design. 3) Planning and re-planning approaches under both dynamic constraints and external environment constraints During executing tasks, UAVs are faced with external environment constraints such as obstacles and other UAVs. It is necessary to calculate the trajectories, satisfying with the external environment constraints and their own dynamic constraints, through the planning method to ensure the safety of the UAVs. In view of this problem, we will first improve the existing kinematic-constraint-based trajectory planning method based on the idea of self-healing control. Then taking advantages of the invariant set: containing system dynamic model and having certain geometry, the invariant set is able to be used as a bridge to connect the dynamic constraints and external environment constraints. Thus, invariant-set-based planning approach is proposed to ensure both of these constraints are satisfied in theory at the same time. On this basis, the problem of re-planning after actuator faults/failures is discussed. 4) Construction of indoor experiment platform and experimental study on rotor failure of a quadrotor UAV Based on a motion capture system, an indoor experiment platform is built to improve the experiment efficiency. Then according to the above theoretical ideas, the performance of quadrotor UAVs under one rotor failure is analyzed. On these basis, two flight experiments are accomplished: a flight of transition mode before and after one rotor failure and another flight of three-dimensional position tracking under one rotor failure.
语种	中文
产权排序	1
内容类型	学位论文
源URL	[http://ir.sia.cn/handle/173321/20557]
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	齐欣. 无人机容错控制与重规划方法研究[D]. 沈阳. 中国科学院沈阳自动化研究所. 2017.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。