Transpose Jacobian-based controllers represent an attractive approach to robot control in Cartesian space. These controllers attempt to drive the robot end-effector posture to a specified desired position and orientation without solving either the inverse kinematics nor computing the robot inverse Jacobian. A wide class of transpose Jacobian-based regulators obtained from the energy shaping plus damping injection technique is analyzed in this paper. Our main theoretical contribution is the introduction of a novel analysis which does not invoke any assumption on Jacobian singularities to ensure local asymptotic stability for a family of nonredundant robots. The performance of four transpose Jacobian-based regulators is illustrated via experimental tests conducted on a direct-drive vertical arm.

Virtual sensors (software abstractions to support programming of sensor use) have been shown to have software-engineering benefits. A sensor integration system is required to support them. We examine the general requirements of such systems and consider the important design requirements. An idealised architecture, ViSIAr, is proposed to serve as a framework for designing and constructing them. Illustrative examples are provided.

In this paper a procedure is proposed for a rational and optimum use of robots. It is the result of the lessons learned by the author in the course of his consultancy and teaching activities on manipulative tasks and programming capabilities of robots.

One of the chief aims of this paper is educational, and has been set on the basis of students' practice with robots in the Laboratory of Robotics at the University of Cassino, Italy. A second goal consists in highlighting and teaching how robots' versatility and flexibility can be easily exploited during robot's operation when a task is properly modelled through elementary actions. An elementary action, representing a small manipulative operation and consisting in one or few simple instructions given in robot language, can be easily performed by the robot's programming capability. Some examples have been reported to better illustrate the ideas and the procedure proposed, whose main advantage, also from an educational viewpoint, consist in providing a simpler method for analysing manipulations for robot design and programming.

Most of the time, the construction of legged robots is made in an empirical way and the optimization of the mechanical structure is seldom taken into account. In order to avoid spending time and money on the construction of many prototypes to test their performance, a CAD tool and a methodology seem to be necessary. In this way it will be possible to optimize on one hand the kinematic structure of the legs, on the other hand the gaits which will be used by the future robot. Thus, we have developed a methodology to design walking structures such as quadrupeds and bipeds, to simulate their dynamic behavior and analyse their performances. The feet/ground interaction is one of the major problem in the context of dynamic simulation for walking devices. Thus, we focus here about the phenomenon of contact. This paper describes a general model for dynamic simulation of contacts between a walking robot and ground. This model considers a force distribution and uses an analytical form for each force depending only on the known state of the robot system. The simulation includes all phenomena that may occur during the locomotion cycle: impact, transition from impact to contact, contact during support with static friction, transition from static to sliding friction, sliding friction and transition from sliding to static friction. Some examples are presented to show the use of this contact model for the simulation of the foot-ground interaction during a walking gait.

Given the high occurence rate in assembly industry, mating a peg into a hole can be considered as one of the most classic problem in robotics. Such a task has been extensively examined by many researchers who have repeatedly attempted to find out a general solution for it. Peg in hole, which is, needlessly to mention, extremely trivial for any human operator, is surprisingly difficult to have it carried out by a robot manipulator. The reason is partly due to a physical limitation of the mechanical compliance of the robot wrist and arm and partly to a lack of a mating strategy allowing the successful execution of the task whatever the initial position of peg and hole axes is. The work presented in this paper tackles a particular class of peg-in-hole (tandem peg-in-hole) and proposes within the behaviour-based paradigm a solution to its two main components (hole search and peg insertion) loosely modeled on the equivalent strategies performed by a blind human being in possess solely of the same sensing capability (i.e. a simple differential touch sensor).

A method of data segmentation, based upon robust least K-th order statistical model fitting (LKS), is proposed and applied to image motion and range data segmentation. The estimation method differs from other approaches using versions of LKS in a number of important ways. Firstly, the value of K is not determined by a complex optimization routine. Secondly, having chosen a fit, the estimation of scale of the noise is not based upon the K-th order statistic of the residuals. Other aspects of the full segmentation scheme include the use of segment contiguity to: (a) reduce the number of random sample fits used in the LKS stage, and (b) to “fill-in” holes caused by isolated miss-classified data.

A six axis robot has been connected to the World Wide Web at http://telerobot.mech.uwa.edu.au/. The robot was made available on the World Wide Web in September 1994 and has been through many revisions since that time. The remote user receives static images of the robot's work space and can manipulate wooden blocks.

The implementation and application of this interface to telerobotics is discussed. Useability and operator behaviour are investigated through analysis of movement records. Questions addressed are: What is the level of interest in this technique and is it a practical interface for telerobotics.

This paper addresses the problem of modeling biped dynamics and the use of such models for the control of walking, running and jumping robots. We describe two approaches to dynamic modeling: the basic Lagrange approach and the non-regular dynamic approach. The new non-regular dynamic approach takes into account discontinuities due to rigid contact between punctual feet and the ground without computing the exact impact time. The contact is close to the physical situation given by non-linear laws (impenetrability, non-smooth contact and real friction cone). Contact dynamics can be well managed with an accurate dynamic model that respects energy consistency during all the phases encountered during a step (0, 1 or 2 contacts). With this model, we can first study the equilibrum of a biped standing on one foot by a linearisation method. In the second stage, the unified modelized equation is used to establish a general control frame based on non-regular dynamical decoupling. A comparison is made and some simulation results are given with a two degree of freedom planar biped robot.

A mathematic model is developed for a quadruped walking robot to investigate the dynamic stability considering all the inertial effects in the system including those of legs. The dynamic model is derived based on Lagrange's equation using matrix-vector notations for the simpler expression. Then an instant gait stability measure is proposed to apply to dynamic gait as well as static gait. The gait stability measure is obtained from the angular momentum of the system about the supporting edges in the quadrupedal gait of +x type. The validity of the gait stability measure is examined along with the gait stability analysis for several representative gait parameters using the developed dynamic model.

This paper proposes a novel hierarchical multi-layer decision tree for representing reactive robot navigation knowledge. In this representation, the perception space is decomposed into a hierarchical set of worlds reflecting environments which are homogeneous in nature and which vary in complexity in an ordered manner. Each world is used to produce a corresponding decision tree which is trained incrementally. The instantaneous perception of the robot is used to select an appropriate rule from the decision tree and a sequence of rule activations form the complete trajectory. The ability to keep the knowledge complexity manageable and under control is an important aspect of the technique.

Cooperating multiple robots may experience slips between their end-effectors and the object when making no rigid grasps, as is the case of manipulating fingers. Consequently, these slips make the robotic system unsuccessful in executing a job properly and can even cause a failure of control. Contact-point slips can be avoided by defining multiple constraints on grasping forces and usable workspace, but these constraints limit the utilization of a robotic system. The researchers working on finger grasps and multiple robots cooperation with contact have always been assuming no slips at the contact points. However, in normal practice, one has to encounter this problem. Hence, this paper has taken this contact-point slip problem as its main topic and theme of work. Consequently, the slip phenomenon is analyzed thoroughly and then generalized for a multiple robot system manipulating a geometrically regular-shaped object. Different slip configurations are introduced and methods for their detection are proposed. This detection does not consider dedicated sensors neither for the object's posture nor for the slip detection. It, rather, utilizes only the finger-tip position information usually available from the kinematics based on very common sensors available with the manipulator. Utilizing the proposed methodology, it is possible to detect the end-point slips in most of the cases and realize an exact task execution and stability of control, with a minimum number of sensors.

During the last two decades a large amount of effort and attention has been paid to the problem of designing and controlling dextrous robotics hands. The possible application background ranges from telerobotics to upper limb prosthetics, while actual industrial developments are mainly limited to specific grippers and tools. Classical problems related to dextrous hand design are kinematics of multifingered hands, development of proper actuation technologies and redundant tendinous systems for transmission. As far as hand control is concerned, grasp preshaping, planning and synthesis are of major concern, while sensor guided hand operation is still a matter of research. The present work reviews the above mentioned hand design and control issues trying to throw some light on the Babel-like confusion encountered when looking at present literature on dextrous hand design and specially control. Most actuation and transmission technologies, as well as control approaches, are studied and classified.

This paper introduces the notion of kinematic redundancy in nonholonomic mechanical systems, identifies some of the interesting properties which result because of the presence of the redundancy, and initiates a study of the control and application of these systems. It is shown that kinematic redundancy in nonholonomic system can be exploited both to simplify the problem of controlling these systems and to enhance their performance capabilities. Moreover, it is demonstrated that these results can be obtained even in the presence of considerable uncertainty regarding the system model. The proposed ideas are illustrated through the study of three example systems: a space robot, a mobile manipulator, and a tractor-trailer system with two steering inputs (fire truck).

In this paper a complete and systematic procedure for the identification of the dynamic parameters of rigid robot manipulators is presented. Starting from the basic results on the subject present in the literature and on a new technique to find exciting trajectories for the estimation, the procedure is developed. A set of algorithms is provided for the implementation of the various steps of the procedure for a generic open-chain structure. The algorithms have been coded in the popular Matlab/Maple environment and the procedure has been tested in a practical case study to identify the dynamic parameters of a six-degree-of-freedom conventional industrial robot.

In this paper, tip tracking control of a single-link flexible robot is studied through computer simulations. The controller investigated is of a two-loop PD type, which is deduced from a control approach originally developed for elastic joint robots. A very simple model, with the flexible link of the robot being lumped to a spring-mass unit, is employed for the controller design. Bounded Input and Bounded Output (BIBO) stable tip tracking performance is guaranteed. The controller is robustly stable under system parameter uncertainties, and is comparatively easy to implement. In the simulations, different payloads are considered, and the system performance is compared with the traditional Singular Perturbation (SP) method. The simulation results show that the PD type controller designed based on the simple lumped model is able to provide quite satisfactory performance.

This paper proposes a navigation system for a non-holonomic mobile robot. The navigation is based on a “look and move” approach. The aim is to define intermediate points called sub-goals through which the robot must pass. This algorithm is particularly suitable for navigation in an unknown environment and obstacle avoidance. Between two successive sub-goals, a shortest path planning solution is adopted. We have adopted the “Dubins' car” because of the environment perception sensor, a 180° laser scanner. In order to minimize the calculation time, the theoretical results of shortest path are approximated by simple equations. The navigation algorithm proposed can be used either in a structured or unstructured environment. In this context the local map construction is based on the segmentation of a structured environment; so for an unstructured environment, a suitable algorithm must be used instead.

In this work a new concept of designing two degree of freedom (2-DOF) planar parallel manipulators (PPMs) is presented. With this design the manipulator's workspace can be increased by increasing the number of cells in the manipulator. A general dynamic model is formulated for the manipulator with any number of cells. The model is adapted for SCARA and ADEPT configurations, and a new approach for balancing these type of manipulators is proposed.

This paper is aimed at presenting a study on the kinematics of the Tricept robot, which comprises a three-degree-of-freedom (dof) parallel structure having a radial link of variable length. The robot workspace is characterized and the inverse kinematics equation is obtained by using spherical coordinates. The inverse differential kinematics and statics are derived in terms of both an analytical and a geometric Jacobian, and a manipulability analysis along the various workspace directions is developed using the concept of force and velocity ellipsoids. A Jacobian-based Closed-Loop Direct Kinematics (CLDK) algorithm is presented to solve the direct kinematics problem along a given trajectory. Simulation results are illustrated for an industrial robot of the Tricept family.

In this paper, a method is proposed for dynamic balancing of hexapods for high-speed applications. The kinematic structure of the hexapod is based on the parallel mechanism. For high-speed applications, hexapod dynamics is the dominant factor, and dynamic balancing becomes very important. The proposed method is aimed at minimizing the changes in the hexapod inertia over the workspace by utilizing the tool holder attached to the hexapod's end-effector as a counterweight.