Planning of an active vision having anthropomorphic features, such as binocularity, foveas and gaze control, is proposed. The aim of the vision is to provide robots with the pose informaton of an adequate object to be grasped by the robots. For this, the paper describes a viewer-oriented fixation point frame and its calibration, active motion and gaze control of the vision, disparity filtering, zoom control, and estimation of the pose of a specific portion of a selected object. On the basis of the importance of the contour information and the scheme of stereo vision in recognizing objects by humans, the occluding contour pairs of objects are used as inputs in order to show the proposed visual planning.

A bending and expanding motion actuator which iscomposed of multi-layered PZT unimorph cells has beendeveloped. These cells are piled one against another.Each cell's electrode is divided into three sectors and thefacing sectors are connected to one another. Multidirectionalbending is realized by controlling the drivingvoltage of each sector. The actuator is 12 mm indiameter, 20 mm in length and 6.4 g in weight with 40cells. The bending angle was 2.2 degrees, and the expansion displacement was 700 μm with 150 V. The resonant frequency was 95 Hz at 25 V.

Force control of walking machines is essential in a natural soft and uneven terrain. A load adaptive PI force control method for the hydraulic actuation system of MECANT I has been developed. The I term of the controller is changed according to the desired load. Force control satisfies the support requirement also when a leg penetrates the terrain. A method to distribute support forces in an optimal manner has been developed. A rule based altitude controller and a dead-zone and saturation based attitude controller calculates the desired body forces. Walking tests on uneven wet sand fields show the useful applicability of the method.

The fundamentals of an approach to solving the control task of robots interaction with a dynamic environment based on the stability of a closed-loop control system are given in this paper. The task is set and solved in its general form. The traditional control concept of compliant robot motion—the hybrid position/force control is discussed. In the paper the proposed control laws ensure simultaneous stabilization of both the desired robot motion and the desired interaction force, as well as their required quality of transient responses. In order to emphasize the fundamental point of this approach in controlling the contact tasks in robotics, the authors have assumed ideal parameters of interacting dynamic systems. The proposed control procedure is demonstrated by one simple example.

This paper describes a compact assembly sequencer for automatic robotic assembly using a configuration space-based path planner. Connection graph of the assembly is constrained by the feasibility information made available from the analysis of local disassembly paths tried by the robot's path planner. Free space of the robot performing assembly is explicitly used to generate the disassembly constraints. We show with the help of a ball pen assembly how the sequencing is performed with this representation.

This paper proposes and demonstrates a method toobtain macroscopic work out of distributed microactuatorsfabricated by IC-compatible micromachiningprocesses. We have coordinated the simple and smallmotion of microactuators in order to perform a task. Theconcept and a control scheme are discussed first. In orderto show the feasibility, the fabrication and operation ofarrayed microactuators for conveyors are described. One uses thermally driven cantilevers and the other uses controlled air flow from micronozzles to carry flat objects.

This paper presents an active visual method for determining the 3-D pose of a flexible object with a hand-eye system. Some simple and effective on-line algorithms to overcome various exceptional situations in chaincoding or determine the object pose more precisely are developed. The pose of a flexible object can be easily changed because of the flexible nature and the prediction of the pose is almost impossible. A new sensing pose is computed by using the image coordinates of the points on the border line of an image window and the current pose of a hand-eye system for the cases that the flexible object is extended outside the window. In a case of exceptional overlapping, a new sensing pose is computed by using the image coordinates of four extreme image points and the current pose of the hand-eye system. Through a chaincoding process on the skeletonized images, the stereo matching problem of two images is transformed into the matching of the curvature representations of the two skeletonized images. The 3-D pose of a flexible object is computed by using the results of this matching and the camera and hand-eye parameters calibrated beforehand. The initial sensing results are used in computing a new sensing pose to determine the object more precisely.

In robotic tasks where the manipulator has to make transition from free space motion to constrained one, there always exists a inevitable impact phase (contact transition phase). Examples can be the autonomous exploratory motion of the end-effector in a cluttered environment, or grasping and manipulation of objects using dexterous mechanical end-effector. A number of controllers have been proposed in the literature with various discussions on their practical implications. In this paper a simple and efficient controller is proposed for the robotic contact tasks. The controller is based on the notion of switching control law where the structure of the controller remains unchanged during the phase transition and only the magnitudes of its gains are changed. The proposed control law maintans the stable performance during the impact phase of the manipulator when making contact with general environment. In addition, the controller exhibits stable and robust performance during the post-contact phase with the desired contact force regulation. The stability of the proposed controller is shown using the Lyapunov method which guarantees the exponential rate of convergence of the state during the contact transition to post-contact phase. Experimental results are presented to demonstrate the practicality of the proposed controller.

In this paper an algorithm is proposed for the problem of path planning of redundant manipulators among obstacles by using a suitable formulation for robot configurations and path strategy. In particular robotic manipulators have been modelled by using reference points on the kinematic chain and their Cartesian coordinates description. The path planning has been formulated as an optimization problem for the determination of adjacent configurations and the path among obstacles with minimum manipulator displacement. The fully Cartesian coordinates description has been useful for the economy of the numerical procedure and for the constraints formulation of link interference and obstacles avoidance constraints. Some examples are reported which prove the practical feasibility of the path planning procedure, and the numerical results have been tested as applicable to industrial robots through easy programming because of the concept of adjacent configurations.

In this paper, we present two DCAL-like (Desired Compensation Adaptation Law) controllers for link position tracking of n-link, rigid, revolute robot manipulators. First, we use a simplified stability analysis to illustrate global asymptotic link position-velocity tracking for a DCAL-like controller with nonlinear feedback. The proof is simplified by employing a different structure for the nonlinear feedback than that originally proposed and by making use of the nonlinear damping control design tool. We then use the nonlinear damping tool to show that a DCAL-like controller with linear feedback can guarantee semi-global asymptotic link position-velocity tracking. The proposed nonlinear and linear feedback DCAL-like controllers are experimetally tested and compared using the Integrated Motion Inc. 2-link direct drive robot manipulator.

A learning control algorithm is proposed for hybrid force/position tracking of a manipulator in contact with a hard surface. The algorithm utilizes acceleration errors and force errors for learning the inputs required to perform repetitive tasks. We prove the convergence of both force and position tracking errors theoretically. The speed of convergence of the tracking errors is shown to be improved by proper choice of learning gains. The method is compared with PI and PID learning, and similarities with the free-space case are outlined. The robustness of the new scheme to uncertainties in surface geometry and link parameters is discussed. Results of simulation studies and experimental implementation on 3-link robot manipulators are presented to illustrate the effectiveness of the proposed learning controller.

Due to the essential development of different means of numerical computation in the last years, new prospects have been opened for realization of different advanced control methods as conventional reasoning, fuzzy rule or ANN-based AI controls. However, it can clearly be seen, that each of these methods have significant technological limits making it expedient to seek compromises between the application of such methods and certain particular hardware solutions designed for a concrete problem. The aim of this paper is to show that in quite wide a range of practically important control tasks appropriate hardware solutions can be elaborated and combined with the above methods.

This paper is addressed at the difficulty of accurately modelling a two-flexible-link manipulator system, which is a necessary pre-requisite for future work developing a high-performance controller for such manipulators. Recent work concerned with the development of an accurate single-flexible-link model is first reviewed and then the expansion of a single-link model into a two-flexible-link system in a way which properly takes into account the coupling and interactions between the two links is discussed. The method of approach taken is to calculate the elastic and rigid motions of the links separately and then to combine these according to the principle of superposition. The application of the model developed is demonstrated in a simulated two-flexiblelink system.

The flexible microactuator (FMA) is a novel pneumaticrubber actuator developed for use in microrobots. Thispaper reports on integrated FMA systems to achievedistributed motion as occurs in intestinal villi, and withmulti-legged arthropods such as centipedes.

For the purpose of miniaturization and integration of FMAs, the authors focussed on two technical issues: A new fabrication process based on stereo lithography and a new FMA design called a restraint beam FMA. Stereo lithography enables fabrication of micro-structures with rubber-like materials suitable for integrated FMAs. The restraint beam FMA makes it possible to fabricate FMAs from a single material, allowing stereo lithography to be used.

As examples of integrated FMA systems, four prototypes are shown: a 5 × 5 FMA array, a 3 × 3 FMA array which has pneumatic circuits at its base, a pipe interior mobile robot, and an amusement system consisting of 30 FMAs, which demonstrates ball handling ability.

Robotic peg-hole insertion operations can play a very important role in manufacturing industry because it is a common requirement in the manufacturing process and, it requires high precision and high speed.

In general, this operation can be divided into two processes: Search process to engage the peg and the hole and, insertion process.

The search process is critical in the assembly operation. It can be defined as a process where the angular.and the translational errors between the peg and the hole are reduced until insertion can occur.

In Part I a technique for the swing-up control of single inverted pendulum system is presented. The requirement is to swing-up a carriage mounted pendulum system from its natural pendent position to its inverted position. It works for all carriage balancing single inverted pendulum systems as the swing-up control algorithm does not require knowledge of the system parameters. Comparison with previous swing-up controls shows that the proposed swing-up control is simpler, eaiser. more efficient, and more robust.

In Part II the technique is extended to the case of the swing-up control of double inverted pendulum systems. Use is made of a novel selective partial-state feedback control law. The nonlinear, open-loop unstable, nonminimum-phase. and interactive MIMO pendulum system is actively linearised and decoupled about a neutrally stable equilibrium by the partial-state feedback control. This technique for swing-up control is not at all sensitive to uncertainties such as modelling error and sensor noise, and is both reliable and robust.

This paper presents the conceptual designs, kinematics and dynamics modeling of a cooperative re-configurable Dual-Arm Cam-Lock Manipulator. A cam-lock manipulator is a robotics structure with a pair of multi-degree of freedom planar arms jointed together at a shared base. This manipulator is designed to be capable of performing a wide variety of tasks by automatically re-configuring itself to form a variable geometry, stiffness, damping, and workspace robotics structure by the virtue of a novel link/joint design along i''s arms, labeled as the cam-lock design. The kinematics and dynamics of this manipulator is described using admissible variables (i.e., variables that define the constrained admissible motion). Along with the dynamic relations of this manipulator, a generic equation was also developed for the joint servo system.

Pattern recognition and object localization, using various sensors such as vision and tactile sensors, are two important areas of research in the application of robotic systems. This paper demonstrates the feasibility of using some relatively inexpensive array of pressure sensors and a neural network approach to achieve object localization and pattern recognition. The sensors that are used are force sensing resistors (FSRs), more specifically, a 16 x 16 array of FSRs. Because of the nonlinearity associated with a FSR, three possible approaches for gathering output from the sensor array are used. The neural network that is used consists of two 2-layer counterpropagation networks (CPNs). One of the CPNs is trained to recognize contact signatures of different objects placed on a fixed reference position on the sensor array.

An approach to time-optimal smooth and collision-free path planning for two industrial robot arms is presented, where path planning and joint trajectory generation are integrated. A suitable objective function, combining the requirements of time optimality and path smoothness, is proposed, which is subject to the continuity of joint trajectories, limits on their rates of change and collision-free constraints. Fast and effective collision detection for the arms is achieved using the Kuhn- Tucker conditions along with the convexity of the distance function and relying on geometrical relationships between cylinders. Nonlinear optimization is used to solve this path planning problem. The feasibility of this method is illustrated both by simulation and by experimental results.

The coupled dynamic response of two cooperating robots handling two flexible payloads is treated using a new algorithm. In this algorithm, the dynamic equations describing the system are obtained using Lagrange's method for the rigid robot links and the finite element method for the flexible payloads. The contact between the flexible payloads is modelled using the penalty function method and a contact search algorithm is employed to identify the contact region.