SIA OpenIR  > 机器人学研究室
四旋翼飞行器自主起降控制问题研究
其他题名Research on Control of Quad-rotor Autonomous Taking off and Landing
周楹君1,2
导师韩建达
分类号TP273
关键词四旋翼 自主起降控制 最优轨迹规划 自动微分 非线性估计
索取号TP273/Z78/2013
页数72页
学位专业模式识别与智能系统
学位名称硕士
2013-05-28
学位授予单位中国科学院沈阳自动化研究所
学位授予地点沈阳
作者部门机器人学研究室
摘要由于具有独特的结构布局和飞行方式,四旋翼飞行机器人能够垂直起降、悬停以及快速转变航向,并且具有良好的敏捷性和操控性,所有这些优秀特性的存在,使得四旋翼飞行机器人系统迅速受到了研究人员的高度重视,被认为在反恐、环境监控等领域具有广阔的应用前景。四旋翼飞行机器人本质上是一种旋翼类飞行机器人,其基本飞行原理是依靠旋翼桨叶高速旋转产生空气流体与机体之间的相对运动,从而利用空气与旋翼之间的作用力与反作用力支撑飞行机器人本身的各种运动。而刻画这种作用力随旋翼转速、机体状态、外部条件等因素变化规律的就是所谓空气动力学模型。由于空气动力学模型中容易受到多种未知因素和外界条件(天气、气流等)的影响,四旋翼飞行机器人动力学模型存在着高度的非线性和不确定性,在自主起降阶段,由于地面效应的存在,这种非线性和不确定性将更为突出和明显,这就为四旋翼飞行机器人的自主控制,尤其是其自主起降控制带来诸多挑战。因此,本论文针对四旋翼飞行机器人自主起降问题,对其中的规划、控制与估计等进行了深入、系统的研究,具体的内容及成果如下:本文首先介绍了四旋翼飞行机器人的结构及基本飞行原理,详细推导并建立了四旋翼飞行机器人的非线性动力学模型,并采用输入输出反馈线性化方法对该数学模型反馈线性化,作为后续自主起降最优轨迹规划和控制的依据。接着研究了自主起降过程的轨迹规划方法。针对当前的规划方法存在的无法充分考虑动力学特性的问题,以运动学规划(Dubins模型)方法为基础,利用四旋翼飞行机器人系统可反馈线性化的特点,提出了一种基于AD算法的全状态信息和控制输入信息的反解方法,从而为优化轨迹的安全评判及优化参数调节提供了依据,增加了自主起降优化轨迹的安全性。随后,为了解决多种不确定因素对控制性能的影响问题,开展了非线性滤波/估计方法研究。针对扩展卡尔曼滤波、无色卡尔曼滤波以及基于MIT规则的自适应无迹卡尔曼滤波等方法进行了比较研究,并在多飞行机器人实验平台上进行实验研究,验证其在实际飞行机器人系统上的可行性和有效性,以为后续的自适应控制策略的设计奠定了基础。最后,在上述研究成果的基础上,研究了自适应自主起降控制方法。以反馈线性化方法为基础,以优化轨迹为期望,通过在线估计外部扰动等不确定因素,提出一种基于估计扰动信息的四旋翼飞行机器人自适应控制方法,并通过仿真验证了方法的有效性。
其他摘要With a unique structure and an unusual way of flight, a Quad-rotor can take off and land vertically, hover, change direction quickly, and have good properties of agility and maneuverability. All these excellent performances make researchers pay high attention to Quad-rotors, which are considered owning vast development prospects in fields of anti-terrorism, environment monitoring and so on. A Quad-rotor is essentially a kind of a rotary-wing helicopter. Its basic principle of flight is utilizing high-speed rotation of rotor blades to generate relative motion between the aircraft body and the air so as to make use of such kind of action and reaction between air and blades to support the motion of a Quad-rotor. And the transformation rule which describes the relation of the rotor speed, the states of the craft, the external conditions and others, is in fact an aerodynamic model. As the model is easy to be influenced by many unknown factors and environmental conditions as weather or airflow, there exist nonlinearity and uncertainty in the dynamic model of a Quad-rotor. And in the stage of autonomous taking off and landing, such nonlinearity and uncertainty appear more clear and obvious because of the ground effect. That has brought lots of challenges for the autonomous control of a Quad-rotor, especially for the autonomous taking off and landing. Therefore, this paper focuses on the autonomous taking off and landing problem of a Quad-rotor and does a deep and scientific research on the planning, controlling and estimation problems. The specific contents and achievements are as follows, This paper first introduces the structure and the basic flight principles of a Quad-rotor, deduces and establishes the nonlinear Quad-rotor dynamic model, and uses input-output feedback linearization to obtain a feedback linearized model, which is the base of the following planning and control of autonomous taking off and landing. Next the trajectories planning methods of autonomous taking off and landing are studied. Considering that there are no methods which take the dynamic characteristics into account, an inverse solution to calculate the whole states and inputs is proposed by using automatic differentiation based on the kinematics analysis (Dubins Model) and the property of feedback linearity. Consequently, this inverse solution offers basis for the safety judgment of the optimized trajectories and for the adjustment of the optimal parameters, which enforces the security for the traces of the autonomous taking off and landing. Then in order to reduce the effects on controlling caused by kinds of uncertainty, nonlinear filtering/estimation methods are studies, special to extended Kalman filter, unscented Kalman filter and MIT based adaptive unscented Kalman filter. Relative experiments were conducted on a multi-flying-robot testbed in order to test and verify the feasibility and validity of the above filters, which lays the foundation of designing an adaptive control strategy. Finally, on the basis of the aforesaid the adaptive control methods for autonomous taking off and landing are discussed. Based on feedback linearization, aimed at tracing an optimal trajectory, an adaptive controller for the Quad-rotor’s autonomous taking off and landing is proposed based on estimating the disturbances on line. And simulation results verify the effectiveness of the proposed controller.
语种中文
产权排序1
文献类型学位论文
条目标识符http://ir.sia.cn/handle/173321/10786
专题机器人学研究室
作者单位1.中国科学院沈阳自动化研究所
2.中国科学院大学
推荐引用方式
GB/T 7714
周楹君. 四旋翼飞行器自主起降控制问题研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2013.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
四旋翼飞行器自主起降控制问题研究.pdf(2355KB) 开放获取CC BY-NC-SA请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[周楹君]的文章
百度学术
百度学术中相似的文章
[周楹君]的文章
必应学术
必应学术中相似的文章
[周楹君]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

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。