In this paper, the variable view imaging system (VVIS) developed to offer flexibility in microscopic observation is applied to microassembly tasks to overcome several problems in normal imaging systems that hindered application of vision-guided micromanipulation techniques such as limited field of view (FOV) and fixed viewing direction. In three representative cases of microassembly, the parts are mated by visual servoing, avoiding FOV problem by combining unit FOV's and occlusion problem by changing the viewing direction to the optimal one. The results demonstrate the superiority and usefulness of the VVIS in various micromanipulation tasks.

Microassembly has become an important technique to fabricate micro devices with different materials, complex shapes, and incompatible processes. Vision-based microassembly is a promising technique for automated microassembly. However, conventional vision-based techniques in microassembly are limited by inherent problems such as a small depth-of-field (DOF) and a narrow field-of-view (FOV). Microassembly operations initially need to detect micro parts in a wide FOV and a large DOF yet also maintain high resolution for the final state. A tradeoff between the DOF (and/or FOV) and resolution prevents the conventional systems from satisfying the aforementioned microassembly requirements. This paper presents an active zooming control method that enables dynamic adjustment of the DOF (FOV) according to the position and the focus measure of micro objects. The proposed method is based on an artificial potential field method with the capability to combine different kinds of constraints such as the FOV, the focus measure, and the joint limits into the system. The stability and robustness of the proposed system are also investigated. The novelty of this method is that it can ensure the vision system maintains a wide FOV and a large DOF initially, and high resolution at the end. Simulations and microassembly experimental results are provided to verify the feasibility of the proposed approach.

Intelligent autonomous mobile robots must be able to sense and recognize 3D environment for navigation and task execution where they live or work. To achieve this, a variety of techniques have been developed for the determination of the 3D geometric information such as stereo vision, laser structured light, laser range finder and so on. However, these methods have several limitations such as susceptibility to noise, low resolution, measurement error in 3D scene. In this paper to overcome these problems we introduce a new sensing method based upon stereo vision methodology which utilizes not only intensity information but also phase information. The novelty of this suggested method comes from the fact that it uses the stereo phase information from the PMP technique together with intensity information captured by stereo image. To verify the efficiency and accuracy of the proposed method, a series of experimental tests is performed. The results show that it greatly enhances correspondence matching performance and measurement resolution.

Intelligent autonomous mobile robots must be able to sense and recognize 3D environment for navigation and task execution, where they live or work. To achieve this, a variety of techniques have been developed for the determination of the 3D scene geometric information such as stereo vision, laser structured light, laser range finder and so on. But these methods have many limitations such as susceptibility to noise, low speed in 3D scene etc. To overcome these limitations we introduce a new sensing algorithm, which is based on the moire technique and stereo vision. To verify the efficiency and accuracy of our algorithm, a series of experimental tests is performed

Intelligent autonomous mobile robots must be able to sense and recognize 3D indoor space where they live or work. In this paper, we propose a new 3D sensing system using the laser structured lighting method, because of its robustness on the nature of the navigation environment and the easy extraction of feature information of interest The proposed active trinocular vision system is composed of a flexible multi-stripe laser projector and two cameras arranged with a triangular shape. Based on the virtual camera concept and the trinocular epipolar constraints, the matching pairs of line features in two real camera images are established, and 3D information from one-shot image can be extracted on the patterned scene. Especially, for robust line feature matching, we propose a new correspondence matching algorithms based on line grouping and probabilistic voting. Finally, a series of experimental tests is performed to show its efficiency and accuracy.

This paper presents a macro/micro positioning system for force control of a chip mounting system. In the proposed macro/micro system, the macro actuator provides the system with gross motion while the micro device yields fine tuned motion to reduce the impact force that occurs between electronic parts and the PCB surface. In order to prove the effectiveness of the proposed macro/micro chip mounting system, we compare the proposed chip mounting head with a conventional head equipped with a macro actuator only. A series of experiments is executed under the mounting conditions of various access velocities and PCB stiffnesses.

Dimensional model fitting finds its applications in various fields of science and engineering and is a relevant subject in computer/machine vision and coordinate metrology. In this paper, we present two new fitting algorithms, distance-based and coordinate-based algorithm, for implicit surfaces and plane curves, which minimize the square sum of the orthogonal error distances between the model feature and the given data points. Each of the two algorithms has its own advantages and is to be purposefully applied to a specific fitting task, considering the implementation and memory space cost, and possibilities of observation weighting. By the new algorithms, the model feature parameters are grouped and simultaneously estimated in terms of form, position, and rotation parameters. The form parameters determine the shape of the model feature and the position/rotation parameters describe the rigid body motion of the model feature. The proposed algorithms are applicable to any kind of implicit surface and plane curve. In this paper, we also describe algorithm implementation and show various examples of orthogonal distance fit.

While navigating, most autonomous mobile robots view things only in front of them and, as a result, they may collide with objects moving from the side or behind. To overcome this problem, an active omni-directional range sensor system is presented, that is capable of obtaining the omni-directional range data on a navigation environment through the use of a laser conic plane and a conic mirror. Based on this system configuration, we propose a self-localization algorithm. The proposed algorithm to estimate the current position and heading angle of mobile robots utilizes the registered range data obtained at two positions; current and previous and matches the two range informations. To show the effectiveness of the proposed algorithm, a series of simulations was conducted under various navigation conditions. The results show that the proposed algorithm is very efficient in processing, and can be effectively utilized for self-localization of mobile robots in unknown environments.

A robotic welding system for closed block assembly in shipyard needs a sensor system for the recognition of the working environments and the weld seam tracking, and a specially designed environment recognition strategy. In the paper, the developed 3D sensor system is briefly introduced, and a strategy of environmental recognition for welding mobile robot navigation is developed in order to recognize work environments efficiently. The task space formed between two lengths within closed blocks is classified into far field. middle field, and near field, according to the robot-to-welding environment distance. The recognition strategy and tactics for sensing the work environment and detecting the obstacles are described and discussed in derail. Finally, a neural network structure for obstacle classification is proposed and tested in a real work environment.

We describe an approach to automation of visual inspection of ball grid array (BGA) solder joint defects of surface mounted components on printed circuit boards by using a neural network. Inherently, the BGA solder joints are located below its own package body, and this induces a difficulty in taking a good image of the solder joints when using a conventional imaging system. To acquire the cross-sectional image of a BGA solder joint, an X-ray cross-sectional imaging method such as laminography and digital tomosynthesis is utilized. However an X-ray cross-sectional image of a BGA solder joint, using laminography or DT methods, has inherent blurring effect and artifact. This problem has been a major obstacle to extracting suitable features for classification. To solve this problem, a neural network based classification method is proposed. The performance of the proposed approach is tested on numerous samples of printed circuit boards and compared with that of a human inspector. Experimental results reveal that the proposed method shows practical usefulness in BGA solder joint inspection.

Automation of welding process in shipyard is ultimately necessary, since welding site is spatially enclosed by floors and girders, and therefore welding operators are exposed to hostile working conditions. To solve this problem, recently a robotic welding system for closed block assembly has been developed. For complete autonomy, this system needs a sensor system for the weld seam tracking and the recognition of the working environments. In this paper, a three-dimensional laser vision system is developed based on the optical triangulation technology in order to provide robots with work environmental map. This sensor system composed of three laser stripes and two visions is used to detect the welding positions and to recognize the 3D shape of the welding environments. The algorithm architecture for the welding environment recognition is divided into two parts. The first is the conventional 3D scanning module, and the second the plane generation module utilizing a Hough transform. The design of the sensor system and the algorithm for sensing the structured environment are described and discussed in detail.

A microsurgical telerobot system has been developed based on the results of the operation task analysis. The telerobot system is composed of a 6-DOF parallel micromanipulator attached to the macro-motion industrial robot and a 6-DOF force-reflecting haptic master device. The master device uses five-bar parallel mechanisms driven by harmonic DC servomotors. The proposed 6-DOF master hand controller has nonlinear coupled dynamics and friction. Since the disturbance force due to friction, gravity and coupled inertia can distort the operator's perception, a disturbance observer is introduced in the operational space and implemented in the microsurgery master hand controller.

Deviation of a slab width from the desired value in hot plate mills has caused significant yield loss by trimming and led to a demand for tighter width tolerances of rolled plates. This necessitates vertical rolling with considerable width accuracy. In this paper, a slab width control system is proposed in order to meet the stringent requirement on the plate dimensional tolerance. The control system adopts a multilayer perceptron neural network to account for the complicated process dynamics characterized by nonlinear, time-varying and uncertain properties. A series of simulation works were conducted to evaluate the performance of the proposed control system for various operating conditions and networks design parameters. The control performance is analyzed in detail in terms of the system response accuracy and robustness to rolling temperature variation.

To obtain good network performance RBF network needs more careful design consideration. The selection of several parameter values such as the number and centers of the radial basis functions must be considered carefully, since they critically affect its performance. We propose a new RBF network architecture, which can recruit neurons automatically to effectively find out appropriate centers best reflecting the characteristics of the input pattern. A self-organizing network, adaptive resonance theory (ART) II is combined with conventional RBF to achieve this. To demonstrate the performance of the proposed network and previously stated effect of network parameters, general problem of function estimation is treated. The representation problem of continuous functions defined over 2D input space is solved. The results obtained from the simulations show that the proposed RBF network yields satisfactory performance in terms of convergence and accuracy compared with those obtained by conventional multilayer perceptron network.

An assembly sequence is considered to be optimal when it minimizes assembly cost while satisfying assembly constraints. The assembly cost relates to assembly operations, assembly motions, and assembly direction changes. This study proposes a simulated-annealing method for the generation of such assembly sequences in robotic assembly. This method reflects the assembly cost to an energy function associated with the assembly sequence. The energy function is iteratively minimized and occasionally perturbed by a simulated annealing until no further change in the energy occurs. As a result, an assembly sequence with a low assembly cost is finally found. To show the effectiveness of the proposed method, case study is presented for an electrical relay. The performance is evaluated from the comparison of the result with the ones obtained from an expert system and a neural network-based approach, previously presented methods.

Unlike rigid parts, deformable parts can be deformed by contact force during assembly. For successful assembly, information about their deformation as well as possible misalignments between the holes and their respective mating parts are essential. In the paper, the locations of marks in three-dimensional space are identified by using a visual sensing system, and part deformation and misalignment in deformable cylindrical peg-in-hole tasks are measured by using the sensing system. Experimental results show that the system and the measurement algorithm are effective in measuring part deformation and misalignment, thereby dramatically increasing the success rate in assembly operations.

This paper presents an active method for locating objects in images, which is conceptualized mainly by elastically deformable contour model. When an initial model is applied to an image data, it attracts near dominant image features such as edges or lines, but tries to keep its home shape or smooth the deformation if a deformation from the home shape occurs. This model is characterized by the core and the rigidity coefficients which control the shape and strength with the prior knowledge about the expected shape of the object. This mechanism significantly improves the performance of detecting object boundaries in presence of some disturbing image features. The proposed method is validated through a series of experiments, which include tracking objects under nonrigid motion and comparison with the original snake models.

A closed-form solution of both forward and reverse position analyses for a 6 degree-of-freedom parallel mechanism of general geometry is derived. The mechanism consists of a top plate, a base, three PPSP serial subchains connecting these two plates. In the position analysis, it is assumed that the first two prismatic joints of each of the three serial subchains are actuated and that the last passive prismatic joint of each of the three subchains are constrained to move in parallel to the top plate, and axes of those three prismatic joints intersect at a common point. Then, the workspace variations of mechanism with respect to joint limitations are investigated. It is concluded that the mechanism studied in this paper has a great advantage over the other parallel mechanism in aspects of analytical simplicity.

Presents simple but effective nonlinear friction compensation methods for a 3-PRPS (prismatic-revolute-prismatic-spherical joints) in-parallel 6-DOF manipulator. It is well known that an integral action in the feedback controller can remove the steady-state error due to constant disturbance. Thus, the integral-action can compensate coulomb friction, which is the DC-component of the total friction and viscous friction in some degrees. When the direction of position command reverses, the integrator output can not change the sign of its output instantaneously, due to friction at zero velocity, i.e. stiction resulting tracking errors. To overcome this problem we attempt to reverse the sign of the integrator output as the sign of velocity changes. The effectiveness of this approach is demonstrated by experiments on a 3-PRPS in-parallel manipulator. Also discussed are disturbance observer and velocity observer approaches for friction compensation.

In this paper, the kinematic analysis for a 6 degree-of-freedom parallel manipulator with 3-PPSP type serial subchains is performed and its kinematic characteristics are investigated via isotropic index of the first-order kinematic influence matrix. Also, dynamic analysis of the manipulator is performed and its dynamic characteristics is examined via the isotropic characteristics of the output effective inertial matrix. From the analysis results, it can be concluded that the manipulator has an excellent kinematic and dynamic characteristics required for high precision manipulators. Lastly, a prototype manipulator system is implemented and the joint position servo-controller is applied to the system. Noting the manipulator has closed-form forward/reverse position solutions, it is expected that more advanced controller requiring heavy computational burden can be applied to the system in real time to further enhance the performance of the system.

In this paper, a new navigation algorithm for wheeled mobile robots is proposed. The proposed algorithm utilizes fuzzy multi-attribute decision-making method in deciding which via-point the robot would proceed at each decision step. Via-point is defined as the local target point of robot's movement at each decision instance. We define a few fuzzy goals to achieve successful robot navigation. At each decision step, a set of the candidates of next via-point in 2D navigation space is constructed by combining various heading angles and velocities. Given the fuzzy goals, the fuzzy decision-making enables the robot to choose the optimal via-point among the candidates. Results from a series of simulations and an experiment in a real environment show the effectiveness of the proposed method.

In this paper, we described an approach to automation of visual inspection of solder joint defects of SMC (surface mounted components) on PCB (printed circuit board) by using neural network and fuzzy rule-based classification method. Inherently, the surface of the solder joints is curved, tiny and specular reflective; it induces a difficulty of taking good image of the solder joints. The shape of the solder joints tends to vary greatly with soldering condition and the shapes are not identical to each other, even for solder joints belonging to a set of the same soldering quality. This problem makes it difficult to classify the solder joints according to their qualities. Neural network and fuzzy rule-based classification method is proposed to efficiently make human-like classification criteria of the solder joint shapes. The performance of the proposed approach is tested on numerous samples of commercial computer PCB boards and compared with the human inspector's performance.

A microsurgical telerobot system has been developed according to the results of the microsurgery task analysis. Several fields of the microsurgery have been investigated and the surgery tasks and tool motions are analyzed to acquire guidelines for specifications of a microsurgical telerobot system. A slave robot system consists of a 6-DOF parallel micromanipulator, that is our own modified Stewart platform type, and a 6-DOF macro-motion industrial robot that mounts the parallel micromanipulator at the end. A 6-DOF force-reflecting master device has also been developed using five-bar parallel mechanisms driven by harmonic DC servomotors. A bilateral control algorithm has been proposed based on the force-position loop, and the local position controller for the master has been developed to minimize the effect of gravity, friction, and inertia.

In this paper, we present the design and the structure of our multiple micro-robot system, CENDORI, designed for playing a robot soccer game, MIROSOT. A robot soccer game is a very complex robot application that incorporates real-time vision, robot control wireless communication and management of multiple robots. To build a robot system with high performance and reliability, we designed the system architecture with a centralized concept. The robot control relies largely upon the vision system and the host computer generating low level motion control words. The experimental results and the performance of the robot are discussed.

This paper presents a kinematic analysis of a parallel manipulator for the probing task application that requires high bandwidth, active compliance, anal high precision. Based on the concept of macro/micro manipulators, the parallel manipulator serves as a wrist of a macro manipulator. The developed manipulator is a class of in-parallel platforms with 3 prismatic-revolute-prismatic-spherical (PRPS) joints chain geometry. The main advantages of this manipulator, compared with the typical Stewart platform type, are the capability of pure rotation generation and the easy prediction of the moving platform motion. The purpose of this paper is to develop an efficient kinematic model which can be used for real-time control and to propose systematic methods to design the manipulator considering the workspace, manipulability, resistivity, and existence conditions of the forward kinematic solution. A series of simulation have been carried out to show the kinematic characteristics and performance of the mechanism. A prototype manipulator has been built based on the kinematic design analysis.

This paper introduces a new vision technique for extracting roof edges of polyhedra having specularly reflecting surfaces. There have been many previous works on object recognition using edge information, but they can not be applied to specular objects since it is hard to acquire reliable camera images of specular objects. If there is a method which can extract the edges of specular objects, it is possible to apply edge-based recognition algorithms to specular objects. To acquire the reliable edge images of specular objects, scanned double pass retroreflection method is proposed, whose main physical characteristic is curvature-sensitive. This utility of the physical characteristic is motivated by the idea that roof edges can be characterized as local surfaces of high curvature. In this paper, the optical characteristics of double pass retroreflection are discussed and a series of simulation studies are performed to verify and analyze the sensor characteristics. The results from a series of simulations show the effectiveness of the proposed method.

This paper presents a kinematic analysis of the parallel manipulator developed for micro-positioning application that requires high control bandwidth and high precision. The manipulator is a class of in-parallel platforms with 3 PRPS (prismatic-revolute-prismatic-spherical) joints chain geometry. The main advantage of this manipulator, compared with the typical Stewart platform type, is the ability to produce pure motion. The purpose of this paper is to develop an efficient kinematic model which can be used for real-time control and to propose systematic design methods that considers the existence conditions of the forward kinematic solution, singularity, manipulability, and resistivity. A series of simulations was carried out to show the kinematic characteristics and performance of the mechanism. The proposed manipulator has potentials in a wide range of applications from micro-surgery robots to a highly dextrous wrist mechanism of assembly robots.

An optical solder joint inspection system (OSJIS) has been developed for the automatic visual inspection of soldered parts on the printed circuit boards. Its advantages over existing techniques include the detection of 3-D shape of specular objects with high reliability and high speed. We propose a solder joint inspection scheme for a prototype of the OSJIS. The inspection scheme is composed of two steps: feature extraction and classification. In the feature extraction step, by scanning a laser beam over the area of solder joints the system obtains two orientation curves representing the quality of soldering condition, and then nine features are extracted from the curves. In the classification step, a neural network classifies the solder joint according to the application requirements by using the features. Experiments were performed for SOP (small outline package)s and QFP (quad flat package)s in insufficient, normal and excess soldering condition. Based upon observation of the experimental results, the proposed inspection scheme shows excellent consistency with visual inspection and a good accuracy of classification performance of 94.2%.

A new probing mechanism and an active compliance control algorithm have been developed for the in-circuit test of a PCB. Commercially available robotic probing devices are incapable of controlling contact force generated when a rigid probe contacts with a solder joint at high speed. This uncontrollable excessive contact force makes some defects on the surface of the solder joint, which is plastically deformable over some limited contact force. This force also makes unstable contact motions resulting in unreliable test data To overcome these problems, the authors propose that serially connected macro and micro actuators with active compliance provide the best potential for the safe and reliable in-circuit test. Moreover, the depth of penetration on a solder joint can be regulated after contact. This paper describes the design characteristics, modeling, and control scheme about the newly proposed devices. The comparison of conventional passive compliance and force feedback control methods with the proposed control scheme for the contact force control is presented. The experimental results clearly show the effectiveness of the proposed system.

In contrast to objects of relatively simple shapes, it is difficult to extract geometrical informations from objects of complicated shapes and thus their assembly requires complete information about a misalignment between mating parts. This motivates development of a new sensing method for detecting the misalignment. In this paper, we propose a new omnidirectional image sensing system for assembly of parts with complicated shapes. It is shown that the system consisting of multiple mirrors and a camera can effectively acquire the coaxial 2n view of the misalignment.

We propose a motion planning method for a mobile manipulator to execute a multiple task which consists of a sequence of point-to-point tasks. In the case of multiple tasks, because the motion trajectories between tasks and the commutation configurations are closely related, we formulate the problem as a global non-linear optimization one which takes into account both the motion trajectories and the commutation configurations. In order to simplify the problem, we divide the problem into two subproblems. At the first step, we obtain the motion trajectories assuming that the commutation configurations are given and at the next step, we obtain the optimal commutation configurations by using genetic algorithms. A series of computer simulation results are demonstrated to verify the effectiveness of the proposed method.

This paper proposes a mathematical approach to the analysis of disassemblability of a product for generating stable robotic assembly sequences. The disassemblability of a part, which represents a cost for disassembling the part from a base assembly, is derived in terms of both separability of a part from a base assembly and motion instability of the base assembly. Then, stable robotic assembly sequences are found in the list of parts, each of which is sequentially chosen to have a locally maximum disassemblability. To show the effectiveness of the proposed method, case studies are presented for industrial products. The results are compared with those of previous researches. The comparison shows that the proposed method can effectively generate stable robotic assembly sequences and provide essential informations on fixture design.

In the paper, we propose a wheel controller with feedforward dynamic compensation. A wheel control system developed for autonomous mobile robots driven by two wheels is described. In the navigation of such mobile robots, large path error occurs mainly due to the disturbance from the floor and the uncompensated dynamics inherent to robot motion. The stable and robust motions in the tracking stage are often difficult to achieve without consideration of wheel dynamics. The feedforward controller is designed based on an approximate inverse model which can be obtained via a series expansion of the plant inverse model. In order to adapt to changes in the environments, the controller is implemented as an adaptive version by employing the online parameter estimator. To demonstrate the validity of the proposed controller, a series of path tracking experiments were conducted for a prototype mobile robot, LCAR, by combining it with a path tracking controller.

The proposed navigator consists of an avoidance behavior and goal-seeking behavior. Two behaviors are independently designed at the design stage and then combined them by a behavior selector at the running stage. A behavior selector using a bistable switching function chooses a behavior at each action step so that the mobile robot can go for the goal position without colliding with obstacles. Fuzzy logic maps the input fuzzy sets representing the mobile robot's state space determined by sensor readings to the output fuzzy sets representing the mobile robot's action space. Fuzzy rule bases are built through the reinforcement learning which requires simple evaluation data rather than thousands of input-output training data. Since the fuzzy rules for each behavior are learned through a reinforcement learning method, the fuzzy rule bases can be easily constructed for more complex environments. In order to find the mobile robot's present state, ultrasonic sensors mounted at the mobile robot are used. The effectiveness of the proposed method is verified by a series of simulations.>

This paper proposes a sensor-based navigation which utilizes multicriteria decision making method in uncertain environments. The multicriteria decision-maker using fuzzy sets is used to select one of avoidance and goal-seeking behaviors according to global mission from higher level planner and situation around the mobile robot. Two behaviors are designed to produce the prescribed responses to currently perceived sensory information using fuzzy logic and reinforcement learning. A decision function to appropriately select one of two behaviors is formulated by use of a fuzzy set. The optimal alternative of two behaviors is defined as that achieving the highest degree of membership in fuzzy decision function. The effectiveness of the proposed method is verified by a series of simulations.>

This paper presents a neural network based computational scheme to generate the optimize robotic assembly sequence for an assembly product consisting of a number of parts. An assembly sequence is considered to be optimal when it meets a number of conditions: it must satisfy assembly constraints, keep the stability of in-process subassembly, and minimize assembly cost. Currently, various search algorithms have been reported for the purpose, but as the number of the parts increases they often fail to generate assembly sequences due to the explosion of the search space. As an alternative solution to overcome this problem, the authors propose a scheme using both the Hopfield neural network and the expert system. Based on the inferred precedence constraints and the assembly costs obtained from the expert system, the authors derive the evolution equation of the network, and finally obtain an optimal assembly sequence resulting from the evolution of the network. To illustrate the suitability of the proposed scheme, case study is presented for an electrical relay. The result is compared with that obtained by the expert system.

A simple and reliable method for measuring the depth and orientation of a local area of an object surface is proposed. The prototype system consists of a camera, a He-Ne laser, and a beam expander for producing a cylindrical beam. The cylindrical beam is actively projected onto the surface of the object, and the distortion of the light pattern is fitted into ellipse. Without computation of 3-D range data, the depth and orientation of object surface are directly determined by ellipse parameters. In a noisy environment, this method reliably offers 3-D information even if the light pattern is partially missing. The measurement error and its source are discussed and experimental results are shown. Due to its simplicity and reliability, this method may be implemented for various robotic tasks requiring 3-D information on the relationship of the robot's end effector to its task environment.

The difficulties in devising reliable assembly strategies result from tlie complexity and the uncertainties of the assembly process and its environments such as imperfect knowledge of the parts being assembled as well as the limitations of the assembly devices performing the assembly. To cope with these problems, we presented a new fuzzy rulebased assembly algorithm for precision parts mating(l7). However, the previous method has limitation in mating precision parts because the parameters in the fuzzy control algorithm were fixed, resulting in rather slow search time. To cope with such limitation, this paper presents a technique which can significantly reduce the search time. The presented technique is based upon the fact that adjustable factors in fuzzy algorithm largely influence the assembly performance. Especially, this investigation focuses on the effect of the output scaling factor. Experimental results show that the proper selection of scaling factors can effectively improve the assembly vagueness associated with actual assembly phenomena. Also, the linguistic fiizzy rules desc.ribed by IF-THEN not only can overcome the diffficulties caused by analytic models but also can effectively construct human thinking and knowledge. However, the previous presented technique had a critical limitation in that the technique requires too much searchin effort for the case of corrective motion generated from the fuzzy rules. Moreover, although control performance was lar ely affected by some fuzzy partitioning and the shape of membership function for a primary fuzzy set, the effect of these parameters was not considered. This pa,per, therefore, investigates the effect of scaling factors on assembly performance arid presents a fast searching technique for large initial error based upon the fuzzy rules. This technique consists of two modes: The one is a fine compensation mode whicli is activated in case of small error, while the other is a coarse compensation mode which is activated in case of large error. The above two modes are switched by heuristically adjusting the output scaling factors depending on the magnitude of errors. The performance of the presented method is evaluated through a series of experiments for chamferless peg-in-hole tasks. large initial error(above lmm) because

Presents a new edge relaxation algorithm that enhances the noisy boundary informations in image. The proposed algorithm employs relaxation process that reduces or even eliminates ambiguities of derivative operator response via contextual informations. The contextual informations are the neighborhood patterns of a central edge, which are estimated using fuzzy pattern matching technique. The algorithm is developed on crack edges. Experimental results show that the proposed scheme is effective even for noisy images or low contrast images.>

A neural-net based control algorithm for the tracking control of a two-wheeled mobile robot is considered. The neural network was designed to generate feedforward control input to nullify the error between output and desired trajectory. The neural network controller was combined with local digital PD controllers for each wheel position control. The feedback-error learning methods is used as the learning mechanism such that the learning and control can be performed simultaneously. To evaluate the effectiveness of the proposed control scheme, a series of tracking simulations were performed for the dynamic model of the mobile robot.>