Visual Motor Control of a 7 DOF Robot Manipulator Using Function Decomposition and Sub-Clustering in Configuration Space

This paper deals with real-time implementation of visual-motor control of a 7 degree of freedom (DOF) robot manipulator using self-organized map (SOM) based learning approach. The robot manipulator considered here is a 7 DOF PowerCube manipulator from Amtec Robotics. The primary objective is to reach a target point in the task space using only a single step movement from any arbitrary initial configuration of the robot manipulator. A new clustering algorithm using Kohonen SOM lattice has been proposed that maintains the fidelity of training data. Two different approaches have been proposed to find an inverse kinematic solution without using any orientation feedback. In the first approach, the inverse Jacobian matrices are learnt from the training data using function decomposition. It is shown that function decomposition leads to significant improvement in accuracy of inverse kinematic solution. In the second approach, a concept called sub-clustering in configuration space is suggested to provide multiple solutions for the inverse kinematic problem. Redundancy is resolved at position level using several criteria. A redundant manipulator is dexterous owing to the availability of multiple configurations for a given end-effector position. However, existing visual motor coordination schemes provide only one inverse kinematic solution for every target position even when the manipulator is kinematically redundant. Thus, the second approach provides a learning architecture that can capture redundancy from the training data. The training data are generated using explicit kinematic model of the combined robot manipulator and camera configuration. The training is carried out off-line and the trained network is used on-line to compute the joint angle vector to reach a target position in a single step only. The accuracy attained is better than the current state of art.     

Discriminative deep belief networks for visual data classification

Visual data classification using insufficient labeled data is a well-known hard problem. Semi-supervise learning, which attempts to exploit the unlabeled data in additional to the labeled ones, has attracted much attention in recent years. This paper proposes a novel semi-supervised classifier called discriminative deep belief networks (DDBN). DDBN utilizes a new deep architecture to integrate the abstraction ability of deep belief nets (DBN) and discriminative ability of backpropagation strategy. For unsupervised learning, DDBN inherits the advantage of DBN, which preserves the information well from high-dimensional features space to low-dimensional embedding. For supervised learning, through a well designed objective function, the backpropagation strategy directly optimizes the classification results in training dataset by refining the parameter space. Moreover, we apply DDBN to visual data classification task and observe an important fact that the learning ability of deep architecture is seriously underrated in real-world applications, especially in visual data analysis. The comparative experiments on standard datasets of different types and different scales demonstrate that the proposed algorithm outperforms both representative semi-supervised classifiers and existing deep learning techniques. For visual dataset, we can further improve the DDBN performance with much larger and deeper architecture.     

Control of knowledgeable manufacturing cell with an unreliable agent

Aiming at the task control problems existing in the knowledgeable manufacturing system, the concept of state jump system is proposed to analyze a knowledgeable manufacturing cell with an unreliable agent. The uncertain factors of the knowledgeable manufacturing cell are addressed in the task control model by utilizing a self-learning method of probability distribution parameters of stochastic events. With the state jump system given, the task control problem is greatly simplified that the optimal task control strategy of the manufacturing cell can be obtained by the combination of the uniform technology and the stochastic dynamic programming. The objective function can be stabilized to a certain extent for different initial conditions, which verifies the feasibility of the control strategy. Compared to the random control and maximum control principles, the objective function value of the optimal control strategy in this paper is relatively low, which confirms the validity of the control strategy.     

Human–robot skills transfer interfaces for a flexible surgical robot

In minimally invasive surgery, tools go through narrow openings and manipulate soft organs to perform surgical tasks. There are limitations in current robot-assisted surgical systems due to the rigidity of robot tools. The aim of the STIFF-FLOP European project is to develop a soft robotic arm to perform surgical tasks. The flexibility of the robot allows the surgeon to move within organs to reach remote areas inside the body and perform challenging procedures in laparoscopy. This article addresses the problem of designing learning interfaces enabling the transfer of skills from human demonstration. Robot programming by demonstration encompasses a wide range of learning strategies, from simple mimicking of the demonstrator's actions to the higher level imitation of the underlying intent extracted from the demonstrations. By focusing on this last form, we study the problem of extracting an objective function explaining the demonstrations from an over-specified set of candidate reward functions, and using this information for self-refinement of the skill. In contrast to inverse reinforcement learning strategies that attempt to explain the observations with reward functions defined for the entire task (or a set of pre-defined reward profiles active for different parts of the task), the proposed approach is based on context-dependent reward-weighted learning, where the robot can learn the relevance of candidate objective functions with respect to the current phase of the task or encountered situation. The robot then exploits this information for skills refinement in the policy parameters space. The proposed approach is tested in simulation with a cutting task performed by the STIFF-FLOP flexible robot, using kinesthetic demonstrations from a Barrett WAM manipulator.     

Dynamic visual servo control of a 4-axis joint tool to track image trajectories during machining complex shapes

A large part of the new generation of computer numerical control systems has adopted an architecture based on robotic systems. This architecture improves the implementation of many manufacturing processes in terms of flexibility, efficiency, accuracy and velocity. This paper presents a 4-axis robot tool based on a joint structure whose primary use is to perform complex machining shapes in some non-contact processes. A new dynamic visual controller is proposed in order to control the 4-axis joint structure, where image information is used in the control loop to guide the robot tool in the machining task. In addition, this controller eliminates the chaotic joint behavior which appears during tracking of the quasi-repetitive trajectories required in machining processes. Moreover, this robot tool can be coupled to a manipulator robot in order to form a multi-robot platform for complex manufacturing tasks. Therefore, the robot tool could perform a machining task using a piece grasped from the workspace by a manipulator robot. This manipulator robot could be guided by using visual information given by the robot tool, thereby obtaining an intelligent multi-robot platform controlled by only one camera.     

Autonomous Bayesian Search and Tracking,and its Experimental Validation

We present a technique that uniformly controls a team of autonomous sensor platforms charged with the dual task of searching for and then tracking a moving target within a recursive Bayesian estimation framework. The proposed technique defines the target detectable region, and uniformly formulates observation likelihoods with detection and no-detection events. The unified likelihood function allows the proposed technique to update and maintain the target belief, regardless of the target detectability. For unified search and tracking (SAT), the proposed technique further predicts the belief in a finite-time horizon, and decides control actions by maximizing a unified objective function consisting of local and global measures derived from the predicted belief. Using the objective function, the proposed technique can smoothly change its control actions even during transitions between SAT. The numerical results first show successful SAT by the proposed technique in tests using a sensor platform with different detectability and comparison with conventional searching techniques under different prior knowledge, and then identifies the superiorities of the proposed technique in SAT. The experimental results finally validate the applicability and extendability of the proposed technique via coordinated SAT in a field experiment.     

Hybrid Force and Vision-Based Contour Following of Planar Robots

Manipulators interacted with uncalibrated environments have limited dexterity due to constraints imposed by unknown environments. However, to perform science or industrial operations, it is necessary to be able to position and orient these manipulators on targets in order to accomplish required control tasks. This article describes how one might enhance manipulator dexterity for planar contour following tasks using hybrid force and vision-based control. The proposed control approach can guarantee task precision employing only a single-axis force sensor and an imprecisely calibrated CCD camera whose optical axis is perpendicular to the planar workspace. The goal of the autonomous task is to drive an instrument mounted on the end-effector of a planar robotic manipulator to follow a visually determined planar contour and continue tracking the contour in desired pose, contact force, and speed, all demanding time-varying, with precision. The proposed control architecture is suitable for applications that require simultaneous force and pose control in unknown environments. Our approach is successfully validated in a real task environment by performing experiments with an industrial robotic manipulator.     

A coordination architecture for UUV fleets

This paper presents a modular and expandable architecture, which includes diversified functions and can be applied to heterogeneous fleets of unmanned underwater vehicles (UUVs), to solve the problem of decentralized formation coordination. The architecture is modular and each module is built such that it can solve a precise task using one or more functions. Three functions among them play a key role for the whole architecture: localization, faultless formation control and fault tolerance. The localization function is performed by the use of an adaptive extended Kalman filter (A-EKF) algorithm; the fault-free formation control function is based on a nonlinear decentralized model predictive control (ND-MPC) algorithm; the fault tolerance function is based on a hierarchy graph theory. The novelty of the paper lies in the use of the above mentioned functions as the core of an architecture which is expandable, decentralized and can be applied to a wide range of vehicles.     

空间机器人最优能耗捕获目标的自适应跟踪控制

提出了一种能够引导末端执行器以期望速度跟踪目标的轨迹规划方法。该方法可以实现避障并满足关节限制要求。基于轨迹规划方法,设计了一种利用自由飘浮空间机器人跟踪与捕获章动自旋卫星的自适应控制策略。此外,该控制策略还考虑了最优能耗、测量误差和优化误差。首先,为了使执行器的跟踪误差和机械臂的能耗最小,将空间机器人的控制策略描述为一个关于关节速度、力矩和避障距离的不等式约束优化问题。然后,推导出一个系数为下三角矩阵的显式状态方程,并对目标函数进行解耦和线性化。设计了一种关节速度和力矩分段优化方法去代替传统的凸二次规划方法求解最优问题,这种方法具有较高的计算效率。最后,利用李雅普诺夫稳定性理论验证了所提控制方法的收敛性。     

Control-Based Solution to Inverse Kinematics for Mobile Manipulators Using Penalty Functions

This study offers the solution at the control feedback level to the inverse kinematics problem subject to state equality and inequality constraints for mobile manipulators. Based on the Lyapunov stability theory, a class of controllers generating the mobile manipulator trajectory whose attractor attained in a finite time, fulfills the above state constraints. The problem of both holonomic manipulability enforcement and collision avoidance is solved here based on an exterior penalty function approach which results in continuous mobile manipulator velocities near obstacles. The numerical simulation results carried out for a mobile manipulator consisting of a nonholonomic wheel and a holonomic manipulator of two revolute kinematic pairs, operating in both a constraint-free task space and task space including obstacles, illustrate the performance of the proposed controllers.     

基于状态空间的机械臂轨迹规划

提出一种基于状态空间的机械臂轨迹规划方法,定义并构造了机械臂系统的状态空间,根据内在机构约束与外部环境约束描述出系统状态的可达范围,并给出了任务的可实现条件.对于可实现任务,在状态空间能搜索到任务完成的最优解.如果任务无法完成,则修改系统配置或约束,在新的状态空间确定任务实现的转化条件,并对任务的设计与规划给予指导.研究了障碍约束下两连杆机械臂的点到点任务,实验结果验证了该方法的有效性.     

A Sequential Bi-criteria Search Algorithm for Robot Path Planning in the Box Pushing Problem

The collision-free trajectory planning method subject to control constraints for mobile manipulators is presented. The robot task is to move from the current configuration to a given final position in the workspace. The motions are planned in order to maximise an instantaneous manipulability measure to avoid manipulator singularities. Inequality constraints on state variables i.e. collision avoidance conditions and mechanical constraints are taken into consideration. The collision avoidance is accomplished by local perturbation of the mobile manipulator motion in the obstacles neighbourhood. The fulfilment of mechanical constraints is ensured by using a penalty function approach. The proposed method guarantees satisfying control limitations resulting from capabilities of robot actuators by applying the trajectory scaling approach. Nonholonomic constraints in a Pfaffian form are explicitly incorporated into the control algorithm. A computer example involving a mobile manipulator consisting of nonholonomic platform (2,0) class and 3DOF RPR type holonomic manipulator operating in a three-dimensional task space is also presented.     

Robust trajectories following control of a 2-link robot manipulator via coordinate transformation for manufacturing applications

A common idea concerning trajectory control of robot manipulators is to tackle the motion of the end-effector. According to traditional trajectory designs, a prescribed profile in a work space is first decomposed into independent joint positions such that the success in a contouring task lies with good tracking capability of individual joints. To advance trajectory control precision without relying on high tracking performance, a contour control strategy for a robot manipulator is presented in this paper. Different from the traditional concept of trajectory control, a contour following control strategy is developed via a coordinate transformation scheme. The main advantage of the proposed control architecture is that the final contouring accuracy will not be degraded in case the tracking performance of the robot manipulator is not good enough. Moreover, using a concept of variable structure control theory, a smooth robust control algorithm is realized in the form of proportional control plus an integration term. The robustness of the control algorithm is also demonstrated. A number of experiments are conducted to demonstrate the advantage of the trajectories following control framework and validate the feasibility of the proposed controller.     

Task space-based control of an underwater robotic system for position keeping in ocean currents

For an underwater vehicle-manipulator system, which consists of an underwater vehicle equipped with a manipulator, it is important to regulate the position of the manipulator’s end-effector with respect to a given target position in many interactive operations. This paper presents a task space-based approach for designing a controller that ensures that the end-effector of an underwater vehicle-manipulator system maintains its position in the presence of unknown ocean currents and uncertainties without the explicit use of a disturbance observer. A feedback linearizing control in task coordinates is used, and an extended Kalman filter (EKF) is employed as a state observer. The proposed approach can also be applied to dynamic positioning or controlled weathervaning of a surface ship whose motion is affected by environmental disturbances. To demonstrate the validity and effectiveness of the proposed approach, numerical simulations and experimental tests were carried out and their results are shown.     

GA-pattern matching-based manipulator control system for real-time visual servoing

_—In robotic applications, tasks of picking and placing are the most fundamental ones. Also, for a robot manipulator, the recognition of its working environment is one of the most important issues to do intelligent tasks, since this aptitude enables it to work in a variable environment. This paper presents a new control strategy for robot manipulators, which utilizes visual information to direct the manipulator in its working space, to pick up an object of known shape, but with arbitrary position and orientation. During the search for an object to be picked up, vision-based control by closed-loop feedback, referred to as visual servoing, is performed to obtain the motion control of the manipulator hand. The system employs a genetic algorithm (GA) and a pattern matching technique to explore the search space and exploit the best solutions by this search technique. The control strategy utilizes the found results of GA-pattern matching in every step of GA evolution to direct the manipulator towards the target object. We named this control strategy step-GA-evnlution. This control method can be applied for manipulator real-time visual servoing and solve its path planning problem in real-time, i.e. in order for the manipulator to adapt the execution of the task by visual information during the process execution. Simulations have been performed, using a two-link planar manipulator and three image models, in order to find which one is the best for real-time visual servoing and the results show the effectiveness of the control method.     

An evolutionary Bayesian belief network methodology for optimum management of groundwater contamination

An integrated methodology, based on Bayesian belief network (BBN) and evolutionary multi-objective optimization (EMO), is proposed for combining available evidence to help water managers evaluate implications, including costs and benefits of alternative actions, and suggest best decision pathways under uncertainty. A Bayesian belief network is a probabilistic graphical model that represents a set of variables and their probabilistic relationships, which also captures historical information about these dependencies. In complex applications where the task of defining the network could be difficult, the proposed methodology can be used in validation of the network structure and the parameters of the probabilistic relationship. Furthermore, in decision problems where it is difficult to choose appropriate combinations of interventions, the states of key variables under the full range of management options cannot be analyzed using a Bayesian belief network alone as a decision support tool. The proposed optimization method is used to deal with complexity in learning about actions and probabilities and also to perform inference. The optimization algorithm generates the state variable values which are fed into the Bayesian belief network. It is possible then to calculate the probabilities for all nodes in the network (belief propagation). Once the probabilities of all the linked nodes have been updated, the objective function values are returned to the optimization tool and the process is repeated. The proposed integrated methodology can help in dealing with uncertainties in decision making pertaining to human behavior. It also eliminates the shortcoming of Bayesian belief networks in introducing boundary constraints on probability of state values of the variables. The effectiveness of the proposed methodology is examined in optimum management of groundwater contamination risks for a well field capture zone outside Copenhagen city.     

Collision‐free control of robotic manipulators in the task space

This paper addresses the problem of position control of robotic manipulators in the task space with obstacles. A computationally simple class of task space regulators consisting of a transpose Jacobian controller plus an integral term including the task error and the gradient of a penalty function generated by obstacles is proposed. The Lyapunov stability theory is used to derive the control scheme. Through the use of the exterior penalty function approach, collision avoidance of the robot with obstacles is ensured. The performance of the proposed control strategy is illustrated through computer simulations for a direct‐drive arm of a SCARA type manipulator operating in both an obstacle‐free task space and a task space including obstacles. © 2005 Wiley Periodicals, Inc.     

运动学补偿的宏2微机器人连续轨迹控制研究

一小机械手附在一大机械手末端构成的系统称为宏－微机器人系统,介绍了宏－微机器人连续轨迹控制的新方法,对宏－微机器人的控制在关节空间进行,通过微机械手的快速运动对宏机械手的轨迹误差在线补偿,宏－微机器人轨迹规划离线进行,通过任务放大的方法分解宏－微机器人运动学冗余,仿真和实验证明了方法的有效性。     

Singular perturbation approach to trajectory tracking of flexible robot with joint elasticity

The control problem of a robot manipulator with flexures both in the links and joints was investigated using the singular perturbation technique. Owing to the combined efects of the link and joint jlexibilities, the dynamics of this type of manipulator become more complex and under-actuated leading to a challenging control task. The singular perturbation being a successful technique for solving control problems with under-actuation was exploited to obtain simpler subsystems with two-time-scale separation, thus enabling easier design of subcontrollers. Furthermore, simultaneous tracking and suppression of vibration of the link andjoint of the manipulator is possible by application of the composite controller, i.e. the superposition of both subcontrol actions. In the first instance, the design of a composite controller was based on a computed torque control for slow dynamics and linear-quadratic fast control. Later, to obtain an improved control performance under model uncertainty, the composite control action was achieved using the radial basis function neural network for the slow control and a linear-quadratic fast control. It was confirmed through numerical simulations that the proposed singular perturbation controllers suppress the joint and link vibrations of the manipulator satisfactorily while a perfect trajectory tracking was achieved.