基于能量平衡的蛇形机器人被动蜿蜒步态研究

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	基于能量平衡的蛇形机器人被动蜿蜒步态研究
作者	张丹凤
学位类别	博士
答辩日期	2015-05-23
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	吴成东 ; 李斌
关键词	蛇形机器人被动蜿蜒步态能量平衡补偿力矩连续体模型
其他题名	Study on the Passive Creeping of a Snake-like Robot Based on Energy Balance
学位专业	模式识别与智能系统
中文摘要	蛇形机器人与环境的交互作用力是随运动变化的，同时具有多样性、不可预测性和不易测量等特点。诸多因素导致控制方法从作用力角度引入环境信息具有复杂性和局限性。基于上述原因，本研究将环境作用力作为未知因素。研究目标是控制蛇形机器人在摩擦系数未知的平面上产生满足期望性能的蜿蜒运动。期望性能即运动需求。平面的摩擦系数未知导致步态与运动性能之间的关系未知，因此蛇形机器人完成运动需求存在困难。能量从供给到动能的转化过程形成了蛇形机器人的运动，本文将这个动态平衡过程称为能量平衡，能量供给取决于控制量决定运动是否满足运动需求，能量耗散从整体角度反应蛇形机器人与环境的交互作用。因此，本研究基于能量平衡对未知环境下蛇形机器人的蜿蜒运动展开研究。首先，针对在摩擦系数未知的平面上，蛇形机器人如何根据与环境的交互信息确定蜿蜒步态的问题，即蜿蜒步态生成方法展开研究。以蛇形机器人蜿蜒运动动能作为研究对象。在摩擦系数未知的平面上，关节控制量与运动动能之间的关系未知，蛇形机器人需要不断地调节步态，选择运动动能满足期望动能的步态。步态调节时间的长短即适应速度，决定平均运动动能。因此，提高未知环境下蛇形机器人蜿蜒步态的适应速度具有重要意义。从生物蛇的运动机理可知，适应速度取决于能量从供给到运动动能的转化效率。在能量转化过程分析的基础上，提出基于能量平衡的蛇形机器人被动蜿蜒步态生成方法。该方法的控制量力矩的幅值受能量需求调节，实现蛇形机器人根据不同的环境和不同的运动需求调节步态；同时，该方法以优化控制效率为目标，控制能量流向最佳方向，实现适应速度的提高。实验证明，该方法较已有方法提高了蛇形机器人在未知环境下蜿蜒步态的适应速度。其次，在步态生成方法的基础上，控制蛇形机器人完成不同的运动需求。由于蛇形机器人在实际应用中转向、避障和路径跟踪运动都涉及到方向控制。同时，线速度是位移、动能等性能的基础。因此，本研究针对蛇形机器人方向和速度两个方面展开如下研究。在未知环境下，蛇形机器人在蜿蜒运动过程中侧滑是不可避免的。未知环境和侧滑导致蛇形机器人的关节控制量与运动方向之间的关系具有不确定性。这种不确定性导致现有方向控制方法不可用。因此，在被动蜿蜒步态生成方法的基础上，提出基于力矩补偿的蛇形机器人方向控制方法。该方向控制方法的原理是利用输入能量的对称性调节蛇形机器人的运动方向，实现蛇形机器人在未知平面环境下沿着期望方向运动到达目标点。同时，提出目标点引导策略，结合基于力矩补偿的蛇形机器人方向控制方法，实现蛇形机器人跟踪不同的曲线和直线路径。实验证明，该方法实现了未知环境下蛇形机器人的定点运动和路径跟踪运动。基于能量平衡的被动蜿蜒步态生成方法实现了蛇形机器人运动动能到达期望动能。相应的，可以通过调节期望动能，实现线速度到达期望值。但线速度存在着波动。因此，需针对消除线速度的波动问题展开研究。未知环境和侧滑导致蛇形机器人线速度与控制量之间的关系未知，因此消除线速度波动存在困难。通过结合理想环境和实际环境下速度波动的影响因素，分析在摩擦系数未知的平面上蛇形机器人蜿蜒运动的控制量与运动速度之间的耦合关系获得最优角频率。在最优角频率控制下，蛇形机器人任一方向上的速度分量达到稳定。在仿真研究中，通过对比不同角频率控制下蛇形机器人运动线速度的波动程度，验证最优角频率的有效性。
索取号	TP242/Z31/2015
英文摘要	The interaction force between the snake-like robot and environment is changed, unpredictable and difficult to be measured, therefore, the control method takes the interaction force as the interaction information is complex. Based on the above reasons, the interaction force is unknown in this paper. The purpose of this paper is that the creeping locomotion of the snake-like robot satisfies the expected performance when the friction coefficient of ground is unknown. In this case, the relation between the joint input and the locomotion performance is unknown, so that the locomotion of snake-like robot satisfies the expected performance is difficult. The conversion of the supply energy into the kinetic energy generates the locomotion of the snake-like robot. The conversion process is called energy balance. The energy supply which is decided by the input of joints decides the locomotion performance. The energy dissipation reflects the interaction between the snake-like robot and the environment. In this paper, the creeping locomotion of the snake-like robot on the unknown ground is studied based on the energy balance. First, as to the locomotion of the snake-like robot on the unknown ground, a gait generation method is proposed. The kinetic energy is the elementary performance of the snake-like robot. The unknown environment leads the coupling relation between the joint input and kinetic energy is unknown, so that the gait has to be continuously adjusted until the kinetic energy satisfies the expected kinetic energy. The time of gait adjustment reflects the adaptability speed which impacts the average kinetic energy, therefore, improving the adaptability speed of the snake-like robot is important. According to the movement principle of the biological snake, the adaptability speed is decided by the energy conversion efficiency. Based on the quantitative analysis of the energy transformation, a control method, called passive creeping based on energy balance, is proposed. The amplitude of torque is adjusted by the energy demand, so that the gait of the snake-like robot is adjusted by the environment and locomotion requirement. In experiment, compared with the existing methods, the adaptability speed of the snake-like robot under the proposed method is improved. Second, the direction control method is the basis of the turning locomotion, obstacle avoidance and path tracking locomotion, and the velocity is the basis of the other performances, such as energy and locomotion distance, so that the locomotion direction and velocity of the snake-like robot on the unknown ground are studied as follow. The sideslip of link is inevitable in the passive creeping. The sideslip and unknown environment cause the coupling relation between the joint input and the locomotion direction is uncertain. The uncertain relation leads the existing direction control method is unavailable. The direction control method of the snake-like robot based on the compensation torque is proposed. The locomotion direction is adjusted by the symmetry of input energy. The snake-like robot achieves moving along the expected direction and reaching the target point under the direction control method. Meanwhile, a target-point guidance strategy is proposed to achieve that the snake-like robot follows the desired path. The results of experiment demonstrate that the snake-like robot can reach the target point and track the desired path under the proposed methods. The passive creeping control method achieves that the linear velocity is fluctuating around the expected velocity by adjusting the joint angle on the unknown ground, while the linear velocity is fluctuating. The difficult to eliminate the fluctuation of the velocity is that the relation between the velocity and the joint input is unknown when the link is slipping and the friction coefficient of ground is unknown. To eliminate the fluctuation of velocity, the relationship between the angular frequency and the linear velocity in the unknown plane is built by combining the factors that the resultant force in ideal environment and the sideslip. In simulation, the comparison of the linear velocities under different angular frequencies demonstrates the validity of the optimal angular frequency.
语种	中文
产权排序	1
页码	108页
内容类型	学位论文
源URL	[http://ir.sia.ac.cn/handle/173321/16795]
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	张丹凤. 基于能量平衡的蛇形机器人被动蜿蜒步态研究[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.