On the Experiment Design for Direct Dynamic Parameter Identification of Parallel Robots

Accurate robot dynamic models require the estimation and validation of the dynamic parameters through experiments. To this end, when performing the experiments, the system has to be properly excited so that the unknown parameters can be accurately estimated. The experiment design basically consists of optimizing the trajectory executed by the robot during the experiment. Due to the restricted workspace with parallel robots this task is more challenging than for serial robots; thus, this paper is focused on the experiment design aimed at dynamic parameter identification of parallel robots. Moreover, a multicriteria algorithm is proposed in order to reduce the deficiencies derived from the single-criterion optimization. The results of the identification using trajectories based on a single criterion and the multicriteria approaches are compared, showing that the proposed optimization can be considered as a suitable procedure for designing exciting trajectories for parameter identification.     

基于6参数模型的6R串联机器人运动学参数辨识

为了避免运动学参数误差辨识中存在参数不连续、计算收敛速度慢的现象,基于一种6参数模型,在DH(Denavit-Hartenberg)法建立的杆件坐标系上建立了6R串联机器人的误差模型,并给出了参数转化公式.设计了计算机仿真实验,在存在驱动器、测量仪器随机噪声误差的条件下对比了使用MDH(改进DH)参数误差模型和6参数模型的仿真辨识效果.6参数模型和MDH模型辨识后定位平均误差分别降低了96.1%和52.9%.结果显示6参数模型具有良好的完备性、连续性.6参数模型的误差参数范围可以从制造公差中得出,辨识速度高于MDH模型,通过公差控制参数范围,消除了没有达到极小性要求对误差辨识的影响.应用此方法对一台SR165型机器人进行参数辨识,定位平均误差由2.5 mm降低至0.35 mm.     

模块化机器人的惯性参数辨识研究

提出了一种基于六维力/力矩传感器的模块化机器人惯性参数辨识的方法。首先,通过Newton-Euler方程建立模块化机器人的动力学方程,然后利用基座力旋量平衡原理建立辨识模型对动力学方程中的未知参数进行辨识,最后以德国AMTEC公司生产的PowerCube模块化机器人实体对这种方法进行了实验验证。     

模块化机器人的惯性参数辨识研究

提出了一种基于六维力／力矩传感器的模块化机器人惯性参数辨识的方法。首先，通过Newton-Euler方程建立模块化机器人的动力学方程,然后利用基座力旋量平衡原理建立辨识模型对动力学方程中的未知参数进行辨识，最后以德国AMTEC公司生产的PowerCube模块化机器人实体对这种方法进行了实验验证。     

The SLAM Algorithm for Multiple Robots Based on Parameter Estimation

With the increasing number of feature points of a map, the dimension of systematic observation is added gradually, which leads to the deviation of the volume points from the desired trajectory and significant errors on the state estimation. An Iterative Squared-Root Cubature Kalman Filter (ISR-CKF) algorithm proposed is aimed at improving the SR-CKF algorithm on the simultaneous localization and mapping (SLAM). By introducing the method of iterative updating, the sample points are re-determined by the estimated value and the square root factor, which keeps the distortion small in the highly nonlinear environment and improves the precision further. A robust tracking Square Root Cubature Kalman Filter algorithm (STF-SRCKF-SLAM) is proposed to solve the problem of reduced accuracy in the condition of state change on the SLAM. The algorithm is predicted according to the kinematic model and observation model of the mobile robot at first, and then the algorithm updates itself by spreading the square root of the error covariance matrix directly, which greatly reduces the computational complexity. At the same time, the time-varying fading factor is introduced in the process of forecasting and updating, and the corresponding weight of the data is adjusted in real time to improve the accuracy of multi-robot localization. The results of simulation shows that the algorithm can improve the accuracy of multi-robot pose effectively.     

Iterative Learning Control for Distributed Parameter Systems Based on Non-Collocated Sensors and Actuators

In this paper, an open-loop PD-type iterative learning control (ILC) scheme is first proposed for two kinds of distributed parameter systems (DPSs) which are described by parabolic partial differential equations using non-collocated sensors and actuators. Then, a closed-loop PD-type ILC algorithm is extended to a class of distributed parameter systems with a non-collocated single sensor and m actuators when the initial states of the system exist some errors. Under some given assumptions, the convergence conditions of output errors for the systems can be obtained. Finally, one numerical example for a distributed parameter system with a single sensor and two actuators is presented to illustrate the effectiveness of the proposed ILC schemes.      

基于PSO的工业机器人时间-脉动最优轨迹规划

为提高工业机器人的工作效率,并且保持机器人关节平稳运动,提出一种基于粒子群优化算法的时间-脉动最优轨迹规划方案;通过权重法将多目标优化转化为单目标优化,再运用粒子群优化算法得到时间-脉动最优的运动轨迹;轨迹规划中,采用了关节空间五次非均匀B样条插值法,以确保脉动曲线的连续性;最后以GRB4016工业机器人为研究对象进行仿真实验,结果表明,该方案可以得到较理想的运动轨迹,并验证了方案的有效性。     

面向铣削任务的工业机器人刚度位姿优化

为提高机器人刚度性能,减小铣削加工误差,对搭载铣削执行器的6自由度机器人进行刚度优化.首先,运用虚功原理建立机器人刚度映射模型;其次,设计辨识实验获取关节刚度;再次,以铣削力椭圆平面的各向同性度为优化指标,运用遗传算法对机器人优化位姿进行求解;最后,对比分析机器人位姿优化前后的整体刚度,并进行机器人铣削试验验证位姿优化的有效性,铣削平面度可提升45％.该优化方法可指导串联型工业机器人对大型航天器舱体的铣削加工任务,提高加工质量.     

动载作用下机场刚性道面模态参数识别研究

关于机场道面刚性结构参数识别问题,研究移动荷载作用下机场刚性道面识别系统中,传统弹性、粘弹性解析识别精度的不足,建立了移动荷载作用下弹性支承边界、弹性嵌固边界和实际边界条件下的道面板的运动方程,提出了一种基于变分法、功互等定理的Kelvin粘弹性地基上道面板系统的识别方法。同时根据最小二乘法准则进行反演,拟合实测动挠度和理论动挠度,从而识别出地基反应模量K。经仿真与实测结果可知,改进方法在保证识别速度的前提下可以显著提高识别精度,具有很好的应用价值,为机场刚性道面的设计、参数识别和质量评价等提供理论依据。     

Parallel Parameter Identification in Industrial Biotechnology

Real-valued black-box optimization of badly behaved and not well understood functions is a wide topic in many scientific areas. Possible applications range from maximizing portfolio profits in financial mathematics over efficient training of neuronal networks in computational linguistics to parameter identification of metabolism models in industrial biotechnology. This paper presents a comparison of several global as well as local optimization strategies applied to the task of efficiently identifying free parameters of a metabolic network model. A focus is being set on the ease of adapting these strategies to modern, highly parallel architectures. Finally an outlook on the possible parallel performance is being presented.     

基于改进Faster RCNN的工业机器人分拣系统

传统的分拣作业无法伴随工作环境的变化进行相应的调整,针对此种不足,出现了基于机器视觉的分拣机器人的相关研究,通过将图像处理和特征工程技术引入视觉模块,使得分拣系统能适时的调整.不同于这些方法,本研究基于实验室的工业分拣系统,将深度学习方法应用其中.通过将Faster RCNN检测算法引入视觉模块并对区域提取网络RPN进行相关改进,加快Faster RCNN模型的检测过程,使得该系统满足工业的实时性要求.Faster RCNN作为一种端到端的方法,能自动对输入图像生成更具表达力的特征,对相应目标提取相应特征,这避免了人工设计特征,它的特征自动生成能力使其能适用于各种场景,这提升了工业分拣机器人的环境适应能力.     

Application of a Perturbation Method for Realistic Dynamic Simulation of Industrial Robots

This paper presents the application of a perturbation method for the closed-loop dynamic simulation of a rigid-link manipulator with joint friction. In this method the perturbed motion of the manipulator is modelled as a first-order perturbation of the nominal manipulator motion. A non-linear finite element method is used to formulate the dynamic equations of the manipulator mechanism. In a closed-loop simulation the driving torques are generated by the control system. Friction torques at the actuator joints are introduced at the stage of perturbed dynamics. For a mathematical model of the friction torques we implemented the LuGre friction model that accounts both for the sliding and pre-sliding regime. To illustrate the method, the motion of a six-axes industrial Stäubli robot is simulated. The manipulation task implies transferring a laser spot along a straight line with a trapezoidal velocity profile. The computed trajectory tracking errors are compared with measured values, where in both cases the tip position is computed from the joint angles using a nominal kinematic robot model. It is found that a closed-loop simulation using a non-linear finite element model of this robot is very time-consuming due to the small time step of the discrete controller. Using the perturbation method with the linearised model a substantial reduction of the computer time is achieved without loss of accuracy.     

一种新型机器人自标定装置及其算法

针对传统基于几何约束的机器人自标定装置仅能对局部工作空间内的机器人位型进行标定测量的问题,提出了一种由安装于机器人末端的球心位置测量装置和可移动球杆组成的新型便携式机器人自标定装置,通过利用球面约束和距离约束,可在较大工作空间内对机器人进行标定测量,从而提高标定结果的可靠性.根据可移动球杆的单、双球布置方式,分别建立了基于向量差和距离差的2种机器人自标定模型及其算法.通过采用局部指数积公式并引入位置伴随变换矩阵,简化了2种自标定模型,从而降低了对运动学方程线性化的计算量.最后,对一种6自由度串联机器人进行了仿真实验,实验结果表明2种自标定算法均能够快速收敛,验证了2种算法的有效性和鲁棒性.     

Learning Controllers for Industrial Robots

One of the most significant cost factors in robotics applications is the design and development of real-time robot control software. Control theory helps when linear controllers have to be developed, but it doesn't sufficiently support the generation of non-linear controllers, although in many cases (such as in compliance control), nonlinear control is essential for achieving high performance. This paper discusses how Machine Learning has been applied to the design of (non-)linear controllers. Several alternative function approximators, including Multilayer Perceptrons (MLP), Radial Basis Function Networks (RBFNs), and Fuzzy Controllers are analyzed and compared, leading to the definition of two major families: Open Field Function Approximators and Locally Receptive Field Function Approximators. It is shown that RBFNs and Fuzzy Controllers bear strong similarities, and that both have a symbolic interpretation. This characteristic allows for applying both symbolic and statistic learning algorithms to synthesize the network layout from a set of examples and, possibly, some background knowledge. Three integrated learning algorithms, two of which are original, are described and evaluated on experimental test cases. The first test case is provided by a robot KUKA IR-361 engaged into the peg-into-hole task, whereas the second is represented by a classical prediction task on the Mackey-Glass time series. From the experimental comparison, it appears that both Fuzzy Controllers and RBFNs synthesised from examples are excellent approximators, and that, in practice, they can be even more accurate than MLPs.Institute for Real-Time Computer Systems & Robotics, University of KarlsruheDepartment of Mechanical Engineering, Division PMA, Katholieke Universiteit Leuven     

基于G代码的双工业机器人协调作业的网络控制

在对双工业机器人协调作业网络控制进行研究的基础上,提出了基于TCP和UDP两种传输协议的客户端/服务器控制模式,实现了不同类型指令和数据的有效传输.同时,研究了单工业机器人控制语言.通过对标准G代码进行扩展与重定义,提出了双工业机器人的坐标定义、直线插补、圆弧插补、螺旋线插补及速度控制等指令,为双工业机器人协调的网络控制提供了一种新的语法,从而解决了基于网络的双工业机器人的控制问题.     

Hybrid Dynamic Control Algorithm for Humanoid Robots Based on Reinforcement Learning

In this paper, hybrid integrated dynamic control algorithm for humanoid locomotion mechanism is presented. The proposed structure of controller involves two feedback loops: model-based dynamic controller including impart-force controller and reinforcement learning feedback controller around zero-moment point. The proposed new reinforcement learning algorithm is based on modified version of actor-critic architecture for dynamic reactive compensation. Simulation experiments were carried out in order to validate the proposed control approach.The obtained simulation results served as the basis for a critical evaluation of the controller performance.     

基于连续动态运动基元的移动机器人路径规划

针对二维动态场景下的移动机器人路径规划问题，提出了一种新颖的路径规划方法——连续动态运动基元（continuous dynamic movement primitives， CDMPs）.该方法将传统的单一动态运动基元推广到连续动态运动基元，通过对演示运动轨迹的学习，获得各运动基元的权重序列，利用相位变量的更新，实现对未知动态目标的追踪.该方法克服了移动机器人对环境模型的依赖，解决了动态场景下追踪运动目标和躲避动态障碍物的路径规划问题.最后通过一系列仿真实验，验证了算法的可行性.仿真实验结果表明，对于动态场景下移动机器人路径规划问题， CDMPs算法比传统的DMPs方法在连续性能和规划效率上具有更好的表现.     

基于动力学模型的轮式移动机器人电机控制

针对移动机器人两路电机协同控制问题,提出基于动力学模型的轮式移动机器人电机控制律 （DMMC）．首先推导出质心位置不一定在几何中心的移动机器人运动学模型和动力学模型,并求解出两轮速 度与力矩之间的非线性微分方程．然后,基于两轮速度与力矩间非线性微分方程、电机电气方程和电机机电 方程,推导出移动机器人系统状态方程．最后采用极点配置得到I 型状态反馈控制律．仿真显示,DMMC 法实 现了对输入指令的零稳态误差快速响应．     

基于标记点的机器人抓取目标期望点动态匹配

针对特征点匹配过程中非标准形状目标物提取的特征点具有随机性,无法对目标物上某一点持续提取;且受环境影响,特征点匹配准确率不高等问题,提出一种基于标记点的匹配方法.首先在目标物期望跟踪点附近贴上若干个标记点,再通过YOLOv3对各标记点进行识别框选;然后在提取区域内进行FAST特征点检测并选取Harris响应值最大的N个...     

Drive Train Optimization for Industrial Robots

This paper presents an optimization strategy for finding the trade-offs between cost, lifetime, and performance when designing the drive train, i.e., gearboxes and electric motors, for new robot concepts. The method is illustrated with an example in which the drive trains of two principal axes on a six-axis serial manipulator are designed. Drive train design for industrial robots is a complex task that requires a concurrent design approach. For instance, the mass properties of one motor affect the torque requirements for another, and the method needs to consider several drive trains simultaneously. Since the trajectory has a large impact on the load on the actuators when running a robot, the method also includes the trajectory generation itself in the design loop. It is shown how the design problem can be formalized as an optimization problem. A non-gradient-based optimization algorithm that can handle mixed variable problems is used to solve the highly nonlinear problem. The outcome from an industrial point of view is minimization of cost and the simulataneous balancing of the trade-off between lifetime and performance.