SIA OpenIR  > 机器人学研究室
仿鳗鱼机器人运动控制方法与高效率步态研究
其他题名Modeling and Control of an Eel Robot and its High Efficiency Gait
张安翻1,2
导师马书根 ; 王越超
分类号TP242
关键词仿鳗鱼机器人 曲线路径跟随 动力学建模 运动控制 切向速度跟 踪控制
索取号TP242/Z31/2018
页数123页
学位专业机械电子工程
学位名称博士
2018-05-24
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
摘要本文的研究内容是围绕国家自然科学基金《水下仿生航行器高效高机动运动的基础理论与关键技术》展开的,本研究的目标在于为仿鳗鱼机器人应用于实际流体环境提供高效率步态支持。具体可分为2个子目标:基于仿鳗鱼机器人的运动学和动力学模型,提供可以补偿流体扰动、建模误差、噪声的L1 自适应运动控制方法,以改善鳗鱼机器人的抗扰动性能,实现切向速度跟踪和一般曲线路径跟随;通过分析现有步态的步态参数与切向速度的关系获得高效率步态。(1)基于牛顿-欧拉方法建立仿鳗鱼机器人的运动学和动力学模型。首先,流体的复杂性和机器人模型结构的高度耦合性、强非线性增加了基于模型的控制器设计的难度。本文分别采用两种简化方法简化原非线性模型:近似简化方法和利用参数对称性简化法。第一种近似化简方法针对现有模型无法直接改写成仿射状态空间形式的问题,采用了泰勒级数展开方法对模型进行了合理近似,但引入了模型误差。第二种化简方法针对已有模型无法得到切向速度子动力学模型的问题,定义了一个特殊的非惯性系,利用参数对称性对模型进行约化,并在该非惯性系下建立了二维动力学模型。该简化方法的优点是没有引入模型误差。二维解析动力学模型的研究给机器人运动控制器设计提供了有效理论支撑。然后,为了减少具有多个关节自由度的仿鳗鱼机器人的算法运行时间,基于迭代牛顿-欧拉算法建立了机器人的3 维动力学模型,其欧拉角向量的引入使得算法更加简洁。该模型可以实现多种3D 步态模式的仿真。(2) 基于近似运动学和简化动力学模型的仿鳗鱼机器人抗扰动运动控制方法研究。为了能使仿鳗鱼机器人应用于实际流体环境,要求运动控制算法能够抵抗流体扰动。针对水下流体环境复杂和动力学模型存在模型误差的现状,在近似模型的基础上设计基于欧拉1(Euler one)方法的分段常数自适应控制器,采用该自适应控制方法主要用于解决流体扰动带来的不确定性和补偿近似模型引入的模型误差,从而实现对关节角的自适应跟踪。然后为了能够实现方向角的控制,对自适应控制律进行了修正,最终实现了方向角的跟踪,仿真验证了自适应控制器在三种不同扰动下的有效性,验证了修正的自适应控制器的有效性。针对现有的速度跟踪控制方法无法应对现实环境的扰动,将步态控制器与P 型迭代学习控制相结合,实现了对“多关节”仿鳗鱼机器人的切向速度跟踪控制。该方法的优势是只要模型的一些极限参数,通过任务的重复性对周期性的系统扰动完全消除,对随机扰动也有较强的抑制能力。与优化算法相比,少量的任务重复就可以达到较好的跟踪效果,此外,发现一种“新步态”模式,该步态往往出现在快速游动时。针对现有的仿鳗鱼机器人曲线路径跟随控制存在曲率限制的问题,基于近似运动学模型(假设侧向滑动远小于前进运动),采用了等级曲线方法,设计了任意曲线路径跟随控制器,该控制器不要求曲率限制,不要求初始位置。仿真分析了鳗鱼机器人在不同模块数,不同步态模式,沿着不同曲线路径,不同初始位置时的跟随能力,结果说明了将路径跟随控制器应用到鳗鱼机器人上是有效的。实验研究了直线路径跟随、曲线路径跟随,结果表明机器人能够跟随期望的2D 曲线路径。(3) 仿鳗鱼机器人的高效率步态优化及步态参数分析。为了获得高效率步态,采用运输经济中的效率指标来评价该步态模式是否是高效率的,将功率输入与切向速度的比作为效率评价指标。而直接优化该目标容易导致奇异,本文将该指标分解为两个子目标:功率输入和切向速度。将输入功率加绝对值后取最大值的和作为输入功率,可有效防止瞬时功率过大的情况。利用多目标优化算法NSGA-Ⅱ算法分析水下鳗鱼机器人在三种不同步态模式下的输入功率、切向速度和步态参数的关系。优化结果发现,新步态模式与蜿蜒步态模式、鳗鱼步态模式相比,在同等切向速度时,消耗的功率较小,在同等输入功率条件下,拥有更快的速度。即新步态是高效率的步态。目前对三维步态的分析较少,而为了获得步态参数与切向速度的关系,利用仿真平台对不同的步态幅值、步态相位差分别进行了分析。仿真发现行波步态模式下,当小于一定幅值和相位差时,幅值越大,切向速度越大;相位差越大,切向速度越小。螺旋步态模式下,存在最优相位差,使得切向速度最小。幅值越大,螺旋曲线的半径越大。相位越大,螺旋曲线的圈数越多。
其他摘要This work is supported by National Natural Science Foundation of China "Basic theory and key technology of high efficiency and high maneuver of underwater biomimetic vehicle". The objective of this study is to provide the high efficiency gait for the eel-like robot to be used in real-fluid environments, which can be divided into two subgoals: to improve the anti-disturbance performance of the eel-like robot by providing motion control approaches based on the kinematics and dynamics of an eel-like robot that can compensate for fluid disturbances, modeling errors and noise, and achieve tangential velocity tracking and general curves path following; obtain high-efficiency gaits by analyzing the relationship between the gait parameters of the existing gaits and the tangential velocity. (1) Kinematic and dynamic model based on Newton-Euler method. Due to the complexity of the fluid and the high degree of coupled nature of the robot, it is not conducive to design a controller based on the model, thus the simplified method is needed to simplify or approximate the completely nonlinear model reasonably. Two simplified methods in this paper are used to simplify the original nonlinear model respectively; the approximate reduction method and the parameter symmetry reduction method. The first approximation reduction method aims at the problem that the existing model cannot be directly rewritten into the affine space form. The Taylor series expansion method is used to reasonably approximate the model, but the model error is introduced. The second simplification method aims at the problem that the existing model cannot obtain the tangential sub-dynamic model, defines a special non-inertial frame, uses the parameter symmetry to reduce the model, and establishes the two-dimensional dynamic model in the non-inertial frame. The advantage of this simplified method is that no model errors are introduced. In order to reduce the algorithm running time of the eel-like robot with multiple joint degrees of freedom, a three-dimensional dynamic model of the robot is modeled based on the iterative Newton-Euler method. The introduction of the Euler angle vector makes the algorithm more concise. The model can simulate multiple 3D gaits. (2) Anti disturbance motion control method based on approximate kinematics and simplified dynamic model. To be able to apply the eel-like robot to the real-fluid environments, the motion control algorithm is required to be able to resist fluid disturbances. For the existence of model errors in the underwater fluid environment complex and dynamic model, a piecewise constant adaptive controller based on the Euler one method is designed on the basis of approximate model, which is used for solving the uncertainty caused by the fluid disturbance and compensating the model error introduced by the approximate model, so as to achieve the adaptive tracking of the joint angle. Then, the adaptive control law is corrected in order to realize the control of the orientation angle, and finally the tracking of the orientation angle is achieved. The simulation verified the effectiveness of the adaptive controller under three different disturbances, and verified the effectiveness of the modified adaptive control. The existing speed tracking control method can not deal with the disturbance of the real environment. Combining the gait controller with P-type iterative learning control, the tangential speed tracking control of the "multi-joint" robot is realized. The advantage of this method is that only some limit parameters of the model are needed. The repeatability of the task completely eliminates periodic system disturbances, and it also has a strong ability to suppress random disturbances. Compared with optimization algorithm, small amount of task repetition can achieve better tracking results. In order to solve the problem of curvature limitation for the existing curve-following path control of squid-like robots, based on the approximate kinematics model (assuming that the lateral slip is much smaller than the forward motion), a graded curve method was used to design an arbitrary curve path following controller. It does not require to limit curvature and initial position. The simulations verify the ability of the eel-like robot to follow the different curves in different gait modes from different initial positions, and indicate that it is effective to apply the path following controller to the eel-like robot. Experiments have investigated linear path, curve path following. The results show that the robot can follow the desired 2D curve path. (3) High efficient gait and gait parameter analysis. To get high-efficiency gait, the efficiency index in transportation economy is used to evaluate whether the gait pattern is highly efficient. However, optimizing the criteria directly always leads to singularity. In this paper, the index is decomposed into two sub criteria: power input and tangential velocity. The ratio of power input to tangential velocity is the index of efficiency. The sum of the maximum value of input power plus absolute value is used as input power, which can effectively prevent the instantaneous power excessive. The relationship between input power, tangential velocity and gait parameters in three different gait patterns is analyzed by multi-objective optimization algorithm NSGA-II algorithm. The optimization results show that the new gait pattern consumes less power at the same tangential speed and has a faster speed at the same input power than that of the serpentine gait and the eel-like gait. At present, there are few analysis on three-dimensional gait, where to obtain the relationship between gait parameters and tangential velocity, different gait amplitudes and gait phase shift are separately analyzed by using the simulation platform. The simulation found that the larger the amplitude, the greater the tangential velocity and the larger the phase difference, the smaller the tangential velocity when the amplitude is smaller than a certain amplitude and phase difference in traveling wave gait. In the helix gait pattern, there is an optimal phase shift that minimizes the tangential velocity. The larger the amplitude, the larger the radius of the spiral curve. The larger the phase shift of the gait, the more laps.
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/173321/21809
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
张安翻. 仿鳗鱼机器人运动控制方法与高效率步态研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2018.
条目包含的文件 下载所有文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
仿鳗鱼机器人运动控制方法与高效率步态研究(22758KB)学位论文 开放获取CC BY-NC-SA浏览 下载
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
[张安翻]的文章
百度学术
百度学术中相似的文章
[张安翻]的文章
必应学术
必应学术中相似的文章
[张安翻]的文章
相关权益政策
暂无数据
收藏/分享
文件名: 仿鳗鱼机器人运动控制方法与高效率步态研究.pdf
格式: Adobe PDF
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。