螺旋轮角差动式管内机器人环境适应性研究

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	螺旋轮角差动式管内机器人环境适应性研究
作者	李特
学位类别	博士
答辩日期	2016-05-25
授予单位	中国科学院沈阳自动化研究所
导师	马书根 ; 王越超
关键词	管内机器人螺旋驱动差动机构参数优化运动性能
其他题名	Study on Environmental Adaptability of In-pipe Robot with Differential Screw Roller Angles
学位专业	机械电子工程
中文摘要	本文的研究内容主要是围绕国家自然科学基金项目“两相状态下管内机器人的自适应移动机理与控制策略研究”展开，立足于提高管内机器人的环境适应性，针对一种新型的螺旋轮角差动式管内机器人的机构优化设计及其运动控制方法开展深入的研究。研究内容主要包括三个方面：螺旋驱动式管内机器人解耦构型设计方法研究、机器人结构参数优化方法研究，以及机器人运动控制策略研究，具体内容包括：(1) 在螺旋驱动式管内机器人构型设计的研究中，旨在解决机器人在直弯管管道环境中的构型耦合问题，以提升管内机器人的环境适应能力。在深入研究机器人运动机理的基础上，对基于公理设计理论的解耦构型设计方法展开了研究。首先，针对直弯管管道的环境需求，建立机器人运动的定量模型，并提出了速度调节机理、负载能力调节机理及转向运动机理。运动机理的研究给机器人构型设计提供了有效理论支撑。然后，提出基于公理设计理论的管内机器人解耦构型设计方法。将直弯管管道环境下的螺旋驱动式管内机器人解耦构型设计问题按照设计需求进行分解，并映射到公理设计理论的设计模型中。根据运动机理和独立公理(耦合性判据)约束，提出一种新型的螺旋轮角差动式管内机器人构型。该机器人的构型设计理念实现了移动速度、负载能力和转向运动之间的解耦，使其能够顺利通过直弯管管道环境。最后，在仿真系统中，初步验证了机器人构型设计的有效性。(2) 机器人的结构参数对其综合运动性能有着重要影响，而作业任务及应用环境特性对机器人的综合运动性能具有一定需求。因此，考虑机器人综合运动性能，基于多目标优化思想对结构参数优化方法展开研究。机器人结构参数优化问题可转化为多目标优化问题。首先提出基于机器人动力学模型的运动性能函数，以构建结构参数多目标优化问题的数学模型。螺旋轮角差动式管内机器人是与环境实时作用的无基座、非完整约束系统。利用Routh方程法构建直弯管管道环境下变约束动力学模型。该模型有利于更加准确地反应机器人穿越管道时的动态运动状态。在动力学模型的基础上，提出了自定心性能、负载性能、弯管通过性能、环形凸台越障性能及环形凹槽越障性能函数。运动性能函数直接反应了机器人结构参数(设计变量)与各运动性能指标(子目标函数)之间的定量映射关系，从而确立了结构参数多目标优化问题的数学模型。为了快速求解该多目标优化问题，提出一种基于改进多目标优化粒子群算法的机器人结构参数优化方法。该算法将指数惯性权重调整方法与基于自适应栅格的多目标优化粒子群算法结合，提高了算法的收敛性及收敛速度。利用该方法进行优化求解，进而利用TOPSIS方法对得到Pareto最优解集进行选优决策。基于多目标优化理论和选优决策的参数优化方法能够得到满足任务需求的综合性能最优的机器人结构参数，并丰富了机器人的机构设计优化理论。在机器人构型设计和结构参数优化的基础上，开展管内机器人的机构设计及控制系统的研究。在设计中，充分考虑了机器人的结构对称性和质量均匀分布性。在完成螺旋轮角差动式管内机器人的样机系统后，进行了验证试验。结果表明，该机器人实现了功能设计目标，并基本符合运动性能理论设计要求。(3) 机器人的优化设计使其从机构层面上具备了一定的环境适应能力。本文分别针对直管环境及直弯管变阻力环境提出了相应的控制方法。在直管道环境中，针对管内机器人常用的两种作业方式(定点作业和恒速巡检作业)，分别提出了定点作业能量最优控制策略和恒速巡检作业能量次优控制策略以提升机器人的能量利用率。控制策略充分利用了机器人螺旋轮角可控的特性。首先，基于直流电动机等效模型建立能量函数，得到机器人运动和能量损耗之间的关系。然后，建立直管内动力学模型，并结合传感器反馈信息估计机器人的负载阻力。最后，根据定点作业和恒速巡检作业的特点对能量函数分别施加约束条件，求解得到机器人面对不同负载阻力时的螺旋轮角控制量和电动机控制量。仿真和试验结果表明，面对不同负载阻力环境，定点作业能量最优控制策略使机器人在保证能耗最优的同时，具有尽可能快的移动速度。恒速巡航能量次优控制策略使机器人在保证巡检速度满足要求的同时，能耗尽可能小。相比于螺旋轮角固定式管内机器人的恒速控制策略，本文所提出的能量优化控制策略能够降低能耗，具有更好的环境适应性。
英文摘要	This research, supported by the Natural Science Foundation of China “Research on the adaptive moving principle of an in-pipe robot under two type working states”, mainly deals with the mechanism optimization design and motion control strategies of an in-pipe robot with differential screw roller angles in order to improve the environmental adaptability. It contains the decoupling configuration design method for an in-pipe robot, the optimization method of the robot’s structure parameters, and the motion control strategies. The detailed contents are listed as follows. (1) The research on the configuration design of the in-pipe robot mainly intends to solve the problem of configuration coupling and improve the environmental adaptability. After discussing the mobile mechanism, a decoupling configuration design method is proposed based on the axiomatic design theory. Firstly, considering the shapes of the straight and curved pipes, the quantified model of the robot movement is established. The velocity adjustment mechanism, load capacity adjustment mechanism, and the steering movement mechanism are proposed. The mobile mechanism provides the theoretical support for the configuration design. Then a decoupling configuration design method based on the axiomatic design theory is proposed. Based on the design requirements, the decoupling configuration design problem is divided in the straight and curved pipes, and it is mapped into the design model of the axiomatic design theory. Under the constrains of the mobile mechanism and the Independent Axiom(Coupling criterion), we propose a new in-pipe robot configuration with differential screw roller angles. It realizes the decoupling of the mobile velocity, the load capacity and the steering movement. The proposed robot can pass through the straight and curved pipes. Finally, the validity of the configuration design is proved by the simulations. (2) The structure parameters have important influence on the comprehensive motion performance, and the tasks and application environment characteristics have some certain demands for the comprehensive motion performance. Therefore, the structure parameter optimization method is studied based on the multi-objective optimization technique in consideration of the comprehensive motion performance. The structure parameter optimization problem can be converted into a multi-objective optimization problem. In order to establish the math model, the motion performance functions are firstly proposed based on the dynamic model. The in-pipe robot with differential screw roller angles is a free pedestal and nonholonomic constraint system. The variable constraint dynamic model is built by the method of Routh’s equations. This model can obtain a real and accurate mobile state. Based on the dynamic model, the functions of the self-centering performance, load performance, steering performance, the obstacle-surmounting performance over the circinate boss, and the obstacle-surmounting performance over the circinate groove are proposed. These functions establish the relations between the structure parameters(design variables) and the mobile performance targets(sub-objective function). So the multi-objective optimization model of the structure parameters is established. In order to solve the problem quickly, an improved multi-objective optimization particle swarm method is proposed. This method combines the adjustment method of inertia weight with the multi-objective optimization particle swarm method based on the adaptive grid. It improves the convergence properties and convergence speed. The proposed method is applied to solving the problem. Then the optimal result of the Pareto set is determined by the TOPSIS method. The proposed method can obtain the parameter design result with an optimal comprehensive performance, and enrich the mechanism design optimization theory as well. The mechanism and control system of the robot are developed based on the configuration design and the parameter optimization. The structure symmetry and mass distributivity are considered in the design. Experiments are carried out after the prototype design. The results show that the proposed robot realizes the function design targets and meets the requirements of the motion performance. (3) The robot has some environmental adaptability through the mechanism optimization design. In respect to the varied resistance environments of straight pipes and curved pipes, control methods are proposed. In order to improve the energy efficiency in the straight pipes, two control strategies of energy optimization: an energy optimization of point-approaching task and an energy suboptimization with constant inspecting speed are proposed in consideration of two general kinds of tasks(point-approaching task and inspection with constant traveling speed). The control strategies make full use of the controllable characteristics of the screw roller angles. Firstly, the energy function is built based on the equivalent model of the DC motor. The relation between the movement and the energy consumption can be obtained from the model. Then, the dynamic model in the straight pipe is established, and the load resistance by the model and sensor information is evaluated. Finally, constraints from the tasks are put on the energy function, and the control variables of the screw roller angles and the motor speed in different load resistance are obtained. The simulation and experiment results indicate that in different load resistance environments, the energy optimization control strategy of point-approaching task can make the robot consume the lest energy and move fast; an energy suboptimization control strategy with constant inspecting speed can make the robot move at the desired speed and consume less energy. Compared to the constant speed control strategy of the in-pipe robot with invariable screw roller angles, the proposed strategies can reduce the energy consumption and improve the environment adaptability of the robot.
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.cn/handle/173321/19679]
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	李特. 螺旋轮角差动式管内机器人环境适应性研究[D]. 中国科学院沈阳自动化研究所. 2016.