| 题名 | 基于力控制的工业机器人精密装配研究 |
| 作者 | 吴炳龙 |
| 学位类别 | 博士 |
| 答辩日期 | 2017-06-01 |
| 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 | 曲道奎 |
| 关键词 | 工业机器人 力控制 力/位混合控制 高精密装配 装配参数优化 |
| 其他题名 | Industrial Robot High Precision Assembly Based on force control |
| 学位专业 | 模式识别与智能系统 |
| 中文摘要 | 本文的研究内容是工业机器人在高精密装配领域的应用，依托于辽宁省科技创新重大专项“智能型搬运机器人”。目前工业机器人已经在装配领域有了大量的应用，但这些应用任务更多的是基于位置控制的工业机器人，因此更多局限于对精度要求不是很高的装配任务。对于精度达到几十微米的高精度装配任务，传统基于位置示教的工业机器人则无能为力。随着力传感器技术和力控制技术的成熟，本文采用基于力控制的工业机器人进行高精密装配作业，并以高精密轴孔装配为例，全面研究高精密装配所涉及关键技术问题，提高精密装配效率和质量。其研究内容主要包括三个方面，分别包括力感知，力控制，在线装配参数优化。三个研究内容概述如下： (1)工业机器人对环境的高精密力感知，是实现高精密装配的必备条件。针对工业机器人采集的力信号，提出了一种多参数整体辨识的方法，辨识的结果用于六维力/力矩传感器采样数据的补偿，可极大提高了末端执行器和环境接触时力的采样精度。辨识的参数包括传感器坐标系和机器人第六轴坐标系之间的偏转角、末端执行器重力、末端执行器重心位置、传感器零位漂移，共11个参数值。该方法优点有两个：一是增加了辨识参数的总数，从而提高了力传感器的采集精度；二是对所需要参数，可整体进行辨识，提高了效率。经过理论分析和实验结果，均验证了该方法的正确性和有效性。 (2)针对工业机器人进行高精密装配，提出一种新的力/位混合控制实现策略，与基于该策略下的轴孔装配流程。装配流程包括搜孔、插入、完成三个阶段，在这三个阶段让机器人实时进行姿态的调整，提高装配的成功率。新的力/位混合控制策略是基于伺服速度环实现，与传统基于伺服位置环实现的力控制系统相比较具有更大的系统带宽，控制系统结构简单，方便实现，也可以把速度信号进行积分，转化为伺服位置环控制的实现方式，具有很大的工程应用价值。仿真与实验结果表明，基于速度环实现的力控制能够更好地跟踪更高频率的正弦给定信号, 具有较好的力跟踪性能。采用力/位混合控制方法和螺旋搜孔的方式能够调整好姿态，顺利地找到装配孔，可以很好的完成轴孔装配作业。 为了融合阻抗控制和力控制的优点，本文可以让机器人能够对外界环境的作用力具有很强的适应能力(阻抗性能)，并且能够随时对外界环境施加所需要的力(力的跟踪性能)。控制系统的实现是基于位置控制的工业机器人，系统结构简单。仿真结果表明，机器人不仅具有刚度控制的特点，对外界约束环境有顺从性，同时能够对期望力很好地进行跟踪。 (3)装配参数优化的目的是寻找得到一组最优的装配参数，让工业机器人快速的完成装配，是一个必备的环节，又是一个费时费力的环节。本文提出一种在线的装配参数优化方法：基于正交探索的高斯过程回归替代的贝叶斯优化算法（OE-GPRBOA），以最优化装配时间为目的。该方法不需要人工参与，解放人力，由机器人自主进行参数的优化，找到最优参数，并进行装配。该方法结合了正交试验设计、高斯过程回归和贝叶斯优化三者的优点，可以减少了试验数量和试验时间，能快速高效地学习试验模型，并对试验结果进行很好的预测和优化。由于高斯过程回归可以处理高维数、小样本和非线性等复杂回归问题，所以该在线优化方法可以适用于各种不同类型的装配任务，实验结果表明，相对于GPRBOA，本文提出的方法（OE-GPRBOA）能够更高的成功率，更短的实验次数，找到最优的装配参数组合，该方法具有较大的经济价值。 在优化装配时间的基础上，本文首次考虑了多目标的装配参数优化问题，优化的目标包括了装配插入过程中轴和孔之间的相互作用力，以提高精密装配的整体的装配质量和效率。本文提出了基于高斯过程回归贝叶斯优化的多目标优化方法，通过将多目标优化转化为单目标的方法，即把多个目标加权综合为装配性能指标，为最终优化的总目标。实验结果表明，采用该在线优化的方法得到的参数，既能以较短的时间完成装配，又能让装配时轴和孔之间的作用力较小，使整体装配过程更加顺滑。 |
| 英文摘要 | Research content of this paper is the application of industrial robot in the field of high precision assembly, which is supported by the Liaoning province major science and technology innovation project: “Intelligent transport robot”. The industrial robot has already been applied in the field of assembly, however the assembly tasks are most based on position controlled industrial robot, so the industrial robot are restricted to the assembly tasks which accuracy requirements are not very high. For the high precision assembly task with the accuracy of tens of microns, the traditional industrial robot based on position control is unable to accomplish. With development of force sensor technology and control technology, this paper adopt force controlled industrial robot to finish high precision assembly. The high precision peg-in-hole assembly is studied for example. This paper comprehensive studied and solved the key problem of high precision assembly, and improved the efficiency and quality of assembly. The research mainly includes three aspects: force sensing, force control, online assembly parameter optimization. The three research contents are summarized as follows: (1) High precision force sensing of industrial robots for the environment, is essential to achieve high-precision assembly. For industrial robot force signal measurement, we proposed a multi-parameter overall identification method, the results of identification used to compensate six-axis force/torque sensor sampling data, can greatly improve the measurement accuracy of contact force between robot end-effector and the environment. The parameters need to be identified comprise a deflection angle between the sensor coordinate system and the robot sixth axis coordinate system, the gravity of end-effector, the center of gravity of robot end-effector, the zero drift of sensor, a total of 11 parameters. This method has two advantages: first, it overall consider a variety of factors, improve the measurement accuracy of the contact force; Second, it can overall identify all parameters, improve the identification efficiency. The theoretical analysis and experimental verification proved the correctness and validity of the method. (2) For the high precision assembly of industrial robots，a new force / position hybrid control strategy and peg-in-hole assembly procedure are proposed. The assembly procedure includes three stages: search, insert and complete. The proposed hybrid force/position control system is based on the motors velocity control loop, and has a larger system bandwidth compared to the traditional force control system based on position control loop, the control structure is simple and easy to implement, and it can integrate the velocity signal, and transform into implementation based on the motors position control loop, it has great engineering application value. The simulation and experimental results show that the force/position hybrid control method based on the velocity control loop has better performance to track higher frequency sinusoidal signal than traditional method which is based on the position control loop; The force/position hybrid control method with spiral search hole strategy can adjust posture, and find the assembly hole well, it can finish the peg-in-hole assembly task well. In order to combine the advantages of impedance control and force control, we proposed a new force tracking impedance control method, which has a strong ability to adapt to the external environment at any moments(impedance performance), and can apply force what we need to the external environment(force tracking performance). Implementation of control system is based on the model of position-based industrial robot, and the structure of this control system is simple. The simulation results show that the robot not only have the characteristics of stiffness control which made robot with compliance to the external environmental constraints, but also can tracking the desired force under different environmental constraints. (3) The purpose of assembly parameter optimization is to find a set of optimal parameters of assembly which can make industrial robots assembled quickly. The assembly parameter optimization is an essential process, and it is a time-consuming process. This paper presents an online robotic assembly parameter optimization method, which is called Gaussian Process Regression surrogated Bayesian Optimization Algorithm based on the Orthogonal Exploration (OE-GPRBOA), to optimize assembly time. This method can liberate the labor, does not require artificial participation. The algorithm can optimize the parameters autonomously, finally find the optimal parameters for robotic assembly. For GPR is suitable for processing high dimension, small size of sample and nonlinear complex regression problems, the proposed OE-GPRBOA method can be used for various assembly tasks. Experimental results show that, the proposed OE-GPRBOA method has more efficiency to find the optimal assembly parameters than GPRBOA, this method can generate big economic impact. Based on optimization of assembly time, this paper considered the multi-objective optimization problem of assembly parameters for the first time. The optimization objectives include the interaction force between peg and hole in insertion process, to improve the quality and efficiency of high precision assembly. This paper proposed the multi-objective optimization method based on the Gaussian Process Regression surrogated Bayesian Optimization Algorithm, The multi-objective optimization is transformed into a single objective method, which is weighted and integrated the multiple objectives into the assembly performance index as final optimization goal. The experimental results show that the parameters obtained by the online optimization method can not only complete the assembly in a short time, but also reduce the interaction force between the peg and the hole, make the whole assembly process more smooth. |
| 语种 | 中文 |
| 产权排序 | 1 |
| 内容类型 | 学位论文 |
| 源URL | [http://ir.sia.cn/handle/173321/20508]  |
| 专题 | 沈阳自动化研究所_其他 |
| 推荐引用方式 GB/T 7714 | 吴炳龙. 基于力控制的工业机器人精密装配研究[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。
修改评论