Manipulating Deformable Linear Objects: Fuzzy-Based Active Vibration Damping Skill

Human can handle a deformable object and damp its vibration with recognized skill. However, for an industrial robot, handling a deformable object with acute vibration is often a difficult task. This paper addresses the problem of active damping skill for handling deformable linear objects (DLOs) by using a strategy inspired from human manipulation skills. The strategy is illustrated by several rules, which are explained by a fuzzy and a P controller. A proportional-integral-derivative (PID) controller is also employed to explain the rules as a comparison. The interpretations from controllers are translated into high level commands in a robotic language V+. A standard industrial robot with a force/torque sensor mounted on the wrist was employed to demonstrate the skill. Experimental results showed the fuzzy based damping skill is quite effective and stable even without any previous acknowledge of the deformable linear objects.Category (5)     

Adaptive hybrid control of manipulators on uncertain flexible objects

This paper presents an adaptive hybrid control approach for a robot manipulator to interact with its flexible object. Because of its flexibility, the object dynamics influence the robot's control system, and since it is usually a distributed parameter system, the object dynamics as seen from the robot change when the robot moves. The problem becomes further complicated such that it is difficult to decompose the robot's position and contact force control loops. In this paper, we approximate the object's distributed parameter model into a lumped 'position state-varying' model. Then, by using the well-known nonlinear feedback compensation, we decompose the robot's control space into a position control subspace and object torque control subspace. We design the optimal state feedback for the position control loop and control the robot's contact force through controlling the resultant torque of the object. We use the model-reference simple adaptive control strategy to control the torque control loop. We also study the problem on how to select a reasonable reference model for this control loop. Experiments of a PUMA robot interacting with an aluminum beam show the effectiveness of our approach.     

基于D-H法的农业采摘机器人运动协作控制系统设计

为提升农业采摘机器人运动协作控制性能,降低机器人碰撞概率,利用D-H法优化设计机器人运动协作控制系统。改装位置、力矩以及碰撞传感器设备,优化运动协作控制器与驱动器,调整系统通信模块结构,完成硬件系统的优化。利用D-H法构建农业采摘机器人数学模型,在该模型下,利用传感器设备实现机器人实时位姿的量化描述,通过机器人采摘流程的模拟,分配机器人运动协作任务,从位置和姿态等多个方面,确定运动协作控制目标,经过受力分析求解机器人实际作用力,最终通过控制量的计算,实现农业采摘机器人的运动协作控制功能。通过系统测试实验得出结论：与传统控制系统相比,机器人位置、姿态角和作用力的控制误差分别降低了约40mm、0.2°和1.2N,在优化设计系统控制下,机器人的碰撞次数得到明显降低。     

Manipulating deformable linear objects: Attachable adjustment‐motions for vibration reduction

This paper addresses the problem of handling deformable linear objects in a suitable way to avoid acute vibration. Different types of adjustment‐motions that eliminate vibration of deformable objects and that can be attached to the end of an arbitrary end‐effector's trajectory are presented. For describing the dynamics of deformable linear objects, the finite element method is used to derive the dynamic differential equations. A genetic algorithm is used to find the optimal adjustment motion for each simulation example. Experiments are conducted to verify the presented manipulating method. © 2001 John Wiley & Sons, Inc.     

Exploration of unknown object by active touch of robot hand

This paper proposes a method of exploring the local shape of an unknown object using the force and torque information obtained from active touch. In the first, we present a method to estimate an unknown curvature, using rolling and sliding motion with a force/torque sensor attached to the fingertip of the hand. Then, the normal curvature equation from 2D curvatures is obtained. Finally we present a reconstruction algorithm of local geometry by using a normal curvature equation, which is composed of principal curvatures and principal directions. The method is tested by using a hand-arm system consisting of an industrial robot arm and an anthropomorphic robot hand with 6-axis force/torque sensor. The feasibility of the proposed method is experimentally validated for objects with simple geometries such as cylinder, spheres etc.     

Simulation of football based on PID controller and BP neural network

Wheel robot soccer speed control system using a ball object detection method and PID controller. A control system is based on the object detection system's behavior based on the robot position's orientation to the target position. PIDs are instruments, pressure, speed, and other operational factors used in control, temperature adjustment flow, and industrial control applications. The PID controller uses control loop feedback dynamics to control functional variables and is the most accurate and stable controller. The robot position is held by placing the ball vertically. When the robot's work is perpendicular to the ball, the robot moves with a certain speed controlled by the PIT controller based on the robot's distance and the ball. Standard conditions (standard ball) test results show that the robot can detect the ball material while in the vertical position, whether on the robot's right or left. In the random test that changes direction, the robot can move more dynamically as the ball's change in place.     

Self-organization of Dynamic Object Features Based on Bidirectional Training

This paper presents a method to self-organize object features that describe object dynamics using bidirectional training. The model is composed of a dynamics learning module and a feature extraction module. Recurrent Neural Network with Parametric Bias (RNNPB) is utilized for the dynamics learning module, learning and self-organizing the sequences of robot and object motions. A hierarchical neural network is linked to the input of RNNPB as the feature extraction module for self-organizing object features that describe the object motions. The two modules are simultaneously trained through bidirectional training using image and motion sequences acquired from the robot's active sensing with objects. Experiments are performed with the robot's pushing motion with a variety of objects to generate sliding, falling over, bouncing and rolling motions. The results have shown that the model is capable of self-organizing object dynamics based on the self-organized features.     

基于机器视觉和力反馈的自动装配技术研究

针对工业生产中自动装配技术装配精度不高的问题,提出了一种基于机器视觉和六维力传感器的自动装配控制方法。使用两个单目摄像头对目标进行两次定位,通过改进的NCC匹配算法进行一次定位,通过基于边缘特征的模板匹配进行二次定位,利用六维力传感器获取装配过程中的力与力矩变化情况,并基于反馈的力与力矩提出了直线运动与螺旋线运动两种装配轨迹规划策略。在搭建的机器人平台上对两种策略进行了对比实验,实验结果表明,在轴孔间隙较大时直线运动装配效率较高,但当轴孔间隙小于0. 1 mm时,直线运动的装配效率和成功率均大幅下降,而螺旋线运动的装配时间主要与装配孔直径有关,对于不同轴孔间隙装配表现稳定,并能以较高成功率实现精度0. 05 mm的轴孔装配。     

Intelligent robotic gripper with adaptive grasping force

The on-off control robot gripper is widely employed in pick-and-place operations in Cartesian space for handling hard objects between two positions. Without contact force monitoring, it can not be applied in fragile or soft objects handling. Although, an appropriate grasping force or gripper opening for each target could be searched by trial-and-error process, it needs expensive force/torque sensor or an accurate gripper position controller. It has too expensive and complex control strategy disadvantages for most of industrial applications. In addition, it can not overcome the target slip problem due to mass uncertainty and dynamic factor. Here, an intelligent gripper is designed with embedded distributed control structure for overcoming the uncertainty of object’s mass and soft/hard features. A communication signal is specified to integrate both robot arm and gripper control kernels for executing the robotic position control and gripper force control functions in sequence. An efficient model-free intelligent fuzzy sliding mode control strategy is employed to design the position and force controllers of gripper, respectively. Experimental results of pick-and-place soft and hard objects with grasping force auto-tuning and anti-slip control strategy are shown by pictures to verify the dynamic performance of this distributed control system. The position and force tracking errors are less than 1 mm and 0.1 N, respectively.

     

Predicting Object Dynamics From Visual Images Through Active Sensing Experiences

Prediction of dynamic features is an important task for determining the manipulation strategies of an object. This paper presents a technique for predicting dynamics of objects relative to the robot's motion from visual images. During the training phase, the authors use the recurrent neural network with parametric bias (RNNPB) to self-organize the dynamics of objects manipulated by the robot into the PB space. The acquired PB values, static images of objects and robot motor values are input into a hierarchical neural network to link the images to dynamic features (PB values). The neural network extracts prominent features that each induce object dynamics. For prediction of the motion sequence of an unknown object, the static image of the object and robot motor value are input into the neural network to calculate the PB values. By inputting the PB values into the closed loop RNNPB, the predicted movements of the object relative to the robot motion are calculated recursively. Experiments were conducted with the humanoid robot Robovie-IIs pushing objects at different heights. The results of the experiment predicting the dynamics of target objects proved that the technique is efficient for predicting the dynamics of the objects.     

Navigation system including associative memory

A new adaptive linear robot control system for a robot work cell that can visually track and intercept stationary and moving objects undergoing arbitrary motion anywhere along its predicted trajectory within the robot's workspace is presented in this paper. The proposed system was designed by integrating a stationary monocular CCD camera with off-the-shelf frame grabber and an industrial robot operation into a single application on the MATLAB platform. A combination of the model based object recognition technique and a learning vector quantization network is used for classifying stationary objects without overlapping. The optical flow technique and the MADALINE network are used for determining the target trajectory and generating the predicted robot trajectory based on visual servoing, respectively. The necessity of determining a model of the robot, camera, all the stationary and moving objects, and environment is eliminated. The location and image features of these objects need not be preprogrammed, marked and known before, and any change in a task is possible without changing the robot program. After the learning process on the robot, it is shown that the KUKA robot is capable of tracking and intercepting both stationary and moving objects at an optimal rendezvous point on the conveyor accurately in real-time.     

When does a robot perceive a dynamic object?

Robots operating in a real‐time environment encounter both stationary and moving objects that need to be negotiated using different schemes in general. Motion planning in a dynamic environment entails tracking moving objects and predicting their future positions. However, this requires as a first step the classification of the objects present in the environment as static or dynamic objects, a step that somehow seems to have been overlooked in the literature dealing with navigation in dynamic environments. Presented here are four schemes for perceiving the presence of dynamic objects in the robot's neighborhood. The first approach incorporates a network architecture that classifies the robot's experience of the environment in terms of spatio‐temporal sensor patterns as an experience of a static or dynamic object. The second method detects motion by observing changes in the map of the environment it builds and updates. The remaining two approaches use a strategy for representing the objects in the environment through clusters; inspecting the characteristics of the clusters reveals the dynamic objects. These methods are denoted as STA (spatio‐temporal approach), MBA (model‐based approach), and CBAI and CBAII (cluster‐based approach I and II), respectively. The methods have been tested in environments that contain multiple dynamic objects amidst static ones and their efficacy established. A brief comparison of these approaches in terms of criteria critical for real‐time collision avoidance has also been presented. © 2002 John Wiley & Sons, Inc.     

Dynamic Rolling-Walk Motion by the Limb Mechanism Robot ASTERISK

New dynamic rolling-walk motion for a multi-legged robot with error compensation is proposed. The motion is realized by using the isotropic leg arrangement and the dynamic center of mass control inspired by bipedal robots. By using the preview control of the zero moment point (ZMP) with a cart-table model based on the bipedal robot's technique, the robot's center of mass trajectory is planned for the dynamic motion. The resolved momentum control for manipulating the multi-links robot as a single mass model is also implemented in the system to maintain the stability of the robot. In the new dynamic rolling-walk motion, the robot switches between the two-leg supporting phase and three-leg supporting phase to achieve dynamic motion with the preview control of the ZMP and resolved momentum control as dynamic motion controllers. The authors analyzed the motion and confirmed the feasibility in the Open Dynamics Engine before testing the motion with an actual robot. Due to the difficulties of controlling the ZMP during the two-leg supporting phase, the authors implemented error compensation by using a gyro sensor and compared the results.     

Sensor-based motion planning in three dimensions for a highly redundant snake robot

A strategy is described for real-time motion planning for a highly redundant snake-like robot manipulator operating in a three-dimensional (3D) environment filled with unknown obstacles of arbitrary shape. The robot consists of many (say 30 or 50) links serially connected by universal joints (such a joint allows a 3D rotation of one link relative to the other). The robot's sensors allow it to sense objects in the vicinity of any points of its body. The task is to move the robot's tip point (its head) from its starting position to a specified target position, collision-free for the whole robot's body. To achieve the efficiency necessary for real-time computation, an iterative procedure is proposed which makes use of a unit motion for a single link based on the tractrix curve. This choice also results in automatically achieving a motion that is 'natural' (in that the joint displacements tend to 'die out' in the direction from head to tail) as well as locally optimal.     

A geometric approach to solving the stable workspace of quadruped bionic robot with hand–foot-integrated function

This paper discusses stable workspace of a hand–foot-integrated quadruped walking robot, which is an important issue for stable operation of the robot. This robot was provided with combined structure of parallel and serial mechanisms, whose stable workspace was the subspace of the workspace in which the system was considered stable. The reachable region was formed under structural conditions, while the stable space was formed by the overall conditions of stability which changed with the robot's pose and the mass of grabbed object. In this paper, based on the robot's main structure, the key issues in solving the robot's workspace are discussed in detail, including searching steady conditions of operation of the robot. To research the robot's workspace, working leg's motion curve needed to be solved by kinematics analysis. Due to the redundant drive, it was problematic to deal analytically with the kinematics of the quadruped walking robot. A geometric method of kinematic analysis was proposed as well. Based on the geometric method, the workspace of the robot under varying postures was analyzed by the method of grid partition and in combination with Matlab, VB and Solidworks software programs. An automated computational system of the stable workspace was developed and an example was given to illustrate the whole process in detail. The theory and analysis procedures were also verified by simulation of the robot and its actual grabbing of an object.     

Distributed object transportation on a desired path based on Constrain and Move strategy

In this paper, a distributed strategy to move objects on different arbitrary paths in a 2D plane is proposed and analyzed. This algorithm which is based on Constrain and Move strategy [M.N. Ahmadabadi, E. Nakano, A Constrain and Move approach to distributed object manipulation, IEEE Trans. Robotics Automation 17 (2) (2001) 157], organizes the robots in two groups. The object manipulation task also is decomposed to two different tasks. The task given to one group is control of linear velocity and that assigned to the other group is control of angular velocity of the object. The independence of these tasks makes the design of the distributed architecture of the team possible. To calculate each robot's desired velocity, a simple method using Constrain and Move strategy and robot's local sensors is developed. To prevent small errors in the robot sensory system from affecting the system performance, limited compliance is assumed in robot arms. The basic behaviors of the robots are presented. Moreover, simulation results are given to verify the proposed strategy.     

Fusing force and vision feedback for manipulating deformable objects

This article describes a framework that fuses vision and force feedback for the control of highly deformable objects. Deformable active contours, or snakes, are used to visually observe changes in object shape over time. Finite‐element models of object deformations are used with force feedback to predict expected visually observed deformations. Our approach improves the performance of large, complex deformable contour tracking over traditional computer vision tracking techniques. This same approach of combining deformable active contours with finite‐element material models is modified so that a vision sensor, i.e., a charge‐coupled device (CCD) camera, can be used as a force sensor. By visually tracking changes in contours on the object, material deflections can be transformed into applied stress estimates through finite element modeling. Therefore, internal object stresses due to object manipulation can be visually observed and controlled, thus creating a framework for deformable object manipulation. A pinhole camera model is used to accomplish vision and force sensor feedback assimilation from these two sensing modalities during manipulation. © 2001 John Wiley & Sons, Inc.     

基于高斯牛顿法的可匹配机器人高精度自动抓放料方法

为了提高可匹配机器人抓料和放料精度,降低抓放料过程中的次品率,提出了基于高斯牛顿法的可匹配机器人高精度自动抓放料方法;通过可匹配机器人抓放料工作站,放置工件原料和成品,针对可匹配机器人夹取能力,设计机器人夹具结构,优化爪臂和驱动装置,增加防滑胶套,设计可匹配机器人自动抓料和放料装置;建立可匹配机器人移动坐标系,构建可匹配机器人自动抓放料数学模型,采用高斯牛顿法,根据泰勒级数展开式,迭代求解可匹配机器人自动抓放料数学模型;利用动态机械工具位姿,调整机器人自定义世界坐标系位姿,校正可匹配机器人自动抓放料误差,实现可匹配机器人高精度自动抓放料;实验结果表明,所提方法的平均抓料和放料准确率分别为92％和95％,抓料和放料过程中次品率分别为0.89％和0.64％,能够有效提高抓料和放料精度,降低抓料和放料过程中次品率.     

Design of a six-axis wrist force/moment sensor using FEM and its fabrication for an intelligent robot

This paper describes the design of a six-axis force/moment sensor using FEM (finite element method) and its fabrication. In order to safely grasp an unknown object using an intelligent hand in robot, the hand has to perceive the weight of it. The weight is calculated by forces F_x, F_y, F_z measured from the six-axis wrist force/moment sensor attached to an intelligent robot's hand. And, in order to accurately push and pull an object, forces and moments should be measured. Also, the position of the robot's finger contacted on an object are calculated by forces F_x, F_y and F_z, and moments M_x, M_y and M_z measured from the six-axis wrist force/moment sensor. Therefore, an intelligent robot's hand should get a six-axis wrist force/moment sensor that can measure forces F_x, F_y and F_z, and moments M_x, M_y and M_z simultaneously. The size of the six-axis force/moment sensor for an intelligent robot’ wrist is very important. If its diameter is larger or its thickness (length) is longer, it cannot be mounted in robot's wrist or it will break down under the applied moment M_x or M_y. So, its size is similar to that of the wrist of human being, that is, the diameter is about 60–80 mm and the thickness (length) about 20–40 mm. But the manufactured sensors are not proper in size for the intelligent robot's wrist. Thus, the six-axis force/moment sensor should be developed for the intelligent robot's wrist.In this paper, the structure of a six-axis wrist force/moment sensor was modeled for an intelligent hand in robot newly. And the sensing elements of it were designed by using FEM and were fabricated by attaching strain-gages on the sensing elements. And, the characteristic test of the developed sensor was carried out. The rated outputs from FEM analysis agree well with the results from the experiments. The interference error of the sensor is less than 2.85%.