This paper is part II of a paper published in the previous issue of Robotica. This part proceeds from the assumption that 3D features have been calssified into either a plane, a 2D corner type I or II, or a 3D corner using the Maximum Likelihood Estimator. The location of the 3D features from the results of the Maximum Likelihood Estimation are derived here. Experimental results characterising the ultrasonic sensor and its application to a robot localisation problem are presented in this paper.

It is well known that computed torque robot control is subjected toperformance degradation due to uncertainties in robot model, and application ofneural network (NN) compensation techniques are promising. In this paper weexamine the effectiveness of neural network (NN) as a compensator for thecomplex problem of Cartesian space control. In particular we examine thedifferences in system performance of accurate position control when the same NNcompensator is applied at different locations in the controller structure. It isfound that using NN to modify the reference trajectory to compensate for modeluncertainties is much more effective than the traditional approach of modifyingcontrol input or joint torque/force. To facilitate the analysis, a new NNtraining signal is introduced and used for all cases. The study is also extendedto non-model based Cartesian control problems. Simulation results withthree-link rotary robot are presented and performances of different compensatinglocations are compared.

A new kinematic design of a parallel spherical wrist with actuatorredundancy is presented. A special feature of this parallel manipulator is thearrangement of co-axial actuators which allows unlimited rotation about any axisinside a cone-shaped workspace. A detailed kinematic analysis has shown thatactuator redundancy not only removes singularities but also increases workspacewhile improving dexterity. The structure optimization has been performed with aglobal dexterity criterion. Using a conditioning measure, a comparison with anon-redundant structure of the same type was performed and shows that asignificant improvement in dexterity has been obtained.

This paper deals with the kinematic synthesis of manipulators. A new method based on distributed solving is used to determine the dimensional parameters of a general manipulator which is able to reach a set of given tasks specified by orientation and position. First, a general Distributed Solving Method (DSM) is presented in three steps: the problem statement, the objective functions formulations and the minimum parameters values determination. Then, this method is applied to solve the synthesis of the Denavit and Hartenberg set of parameters of a manipulator with a given kinematic structure. In this case, the kind and the number of joints are specified and a set of constraints are included such as joint limits, range of dimensional parameters and geometrical obstacles avoidance. We show that if the Denavit and Hartenberg parameters (DH) are known, the synthesis problem is reduced to an inverse kinematic problem. We show also how the problem of robot base placement can be solved by the same method. A general algorithm is given for solving the synthesis problem for all kind of manipulators. The main contribution of this paper is a general method for kinematic synthesis of all kind of manipulators and some examples are presented for a six degrees of freedom manipulator in cluttered environment.

This paper presents a novel time-optimal motion planning strategy for a mobile robot with kinematic constraints. The method works in environments in presence of obstacles, without needing to generate the configuration space for the robot. Further, it derives a minimum time first derivative smooth path, as opposed to a minimum distance path which is commonly given by various present solution techniques. The problem is solved in three stages: (i) A reduced visibility graph for a point object is obtained. (ii) The reduced visibility graph is converted into a feasible reduced visibility graph accounting for the size and kinematic constraints of the robot. (iii) The A* algorithm is used to search the feasible reduced visibility graph with the cost function being the time of travel, to obtain a safe, time-optimal, smooth path. The algorithm runs in polynomial time. The method has been tested in computer simulations and test results are presented

In some daily tasks, such as pick and place, the robot is requested to reach with its hand tip a desired target location while it is operating in its environment. Such tasks become more complex in environments cluttered with obstacles, since the constraint for collision-free movement must be also taken into account. This paper presents a new technique based on genetic algorithms (GAs) to solve the path planning problem of articulated redundant robot manipulators. The efficiency of the proposed GA is demonstrated through multiple experiments carried out on several robots with redundant degrees-of-freedom. Finally, the computational complexity of the proposed solution is estimated, in the worst case.

Variable structure control is applied to an underactuated two andthree-link robot. Control laws are developed to invert the robot from itsstable, downward equilibrium position to an inverted position. This maneuver isaccomplished by pumping energy into the system with the variable structurecontroller. A linear balancing controller is activated once the robot nears theinverted position. Numerical simulations are presented for both the two andthree-link robot.

In this paper we present the SMAR CAD-robotics system (Système de Modélisation et d'Animation de Robots), which we developed at the University of Poitiers. This system allows its user to deal with a great number of robotics problems through the use of a graphic simulator. We will discuss the different parts which form the SMAR system. This includes the following:

—The modeler which allows the user to build a database, describing the robot and its environment. The database generated by the system is composed of the geometric description of the objects and the kinematics description of the environment.

—The simulator and the coordinates reverser, which simulate the robot's movements.

—The collision detection algorithms used to verify task accomplishment.

—A calculation algorithm in order to find optimal placement, which determines the relative position robot/task, allowing the robot to efficiently execute the assigned task.

—The collision free-path planning algorithm allowing the system to generate trajectories in a cluttered environment.

An example dealing with a complex robotized cell will also be presented in order to demonstrate the capabilities of the system.

An assembly sequence is considered to be optimal when the sequence satisfies assembly constraints and yields the minimum assembly cost. While, a line balancing solution is considered to be optimal when the solution has the minimum idle time of the line, i.e. the minimum number of workstations for a given cycle time. Although optimal assembly sequences are generated without considering line balancing, they may not guarantee the minimum number of workstations. In such case, additional cost should be required to accommodate the increased number of workstations. Therefore, it is essential to consider line balancing in the generation of cost-effective assembly sequences. To generate such line-balanced assembly sequences for robotic assembly, this paper treats a single-model and deterministic (SMD) assembly line balancing (ALB) problem, and proposes a new method using a simulated annealing. In this method, an energy function is derived in consideration of the satisfaction of assembly constraints, and the minimization of assembly cost and the idle time. Then, the energy function is iteratively minimized and occasionally perturbed by a simulated annealing. When no further change in energy occurs, a solution of assembly sequence with consideration of line balancing is finally obtained. To show the effectiveness of the proposed scheme, case studies are presented for industrial products such as an electrical relay and an automobile alternator.

In this paper the minimum cost trajectory planning problem with fixed time in robot manipulators is considered. The task is solved by transforming the problem to a set of free right-end time optimal problems, leading to a suboptimal solution. Each problem of the optimal cost trajectory planning with a free time is effectively solved by the method of minimal neighbourhood. An algorithm for the task of suboptimal cost trajectory planning with fixed time is presented and applied to the model of a PUMA-like robot. Results of the paper seem to be of particular relevance to the optimization of multi-robot systems.

In this paper, we will present a new 6-DOF parallel robot using a set oftwo Delta structures. An effective method is proposed to establish explicitrelationships between the end effector co-ordinates and the active and passivejoint variables. A simulation of the 2-Delta robot on a C.A.D. Robotics systemwill also be presented. This simulation will allow us to validate the cohesionof our calculations, and to show the workspace depending on the mechanicallimits on passive joints variables. Finally, an approach is proposed to studythe influence of small clearances of the passive joint on the precision of theposition and rotation of the effector. This approach is based on a conceptsimilar to that of Yoshikawa's manipulability.

A laser range finder mounted on a site and azimuth turret is used as a 3D range camera. It forms, associated with a video camera, an original stereovision system. The internal structure of both images are the same but the resolution of 3D image stays low. By ignoring the acquiring speed of measures, spatial resolution is limited by the accuracy of deviation device and the laser footprint. The fact that the impact of the beam is not a point introduces spatial integration.

To correct the average at depth discontinuities due to the beam footprint, a neural-network-based solution is reported.

The use of such a multisensor system requires its calibration. As camera calibration is a well-known problem, the paper focuses on models and calibration methods of the range finder. Experimental results illustrate the quality of the calibration step in terms of accuracy and stability.

The footprint correction is evaluated for both 1D and 2D range finder scannings.

The first stage in developing a task planning language for redundant manipulators is that of constructing an algorithm to produce a schedule of joint displacements for the specified task. This algorithm may be based upon the use of an inverse infinitesimal kinematic analysis for the majority of the path with a finite inverse analysis used to provide an initial joint configuration and periodically to correct for any joint errors that may have accumulated. Here, it is assumed that there is complete correlation and coherence of solutions between these analyses, otherwise problems, such as a lack of repeatability may occur. This may be accomplished by modelling the manipulator as an elastic system which permits the development of a task planning method by using analyses which are all mutually coherent. In this paper, the method is developed and demonstrated using a redundant manipulator that simultaneously avoids obstables, singularities and joint motion limitations. The task planning method and associated analyses are applicable to the general serial redundant manipulator with an arbitrary degree of redundancy.

Isotropic velocity radius (IVR) and isotropic accelerationradius (IVR) are proposed as local performance indices to quantify thedynamic responsiveness of a multi-arm robot as regards the velocity effects.These performance measures are defined on the basis of the acceleration setdescribing the effects of actuator torques, velocity of the manipulated object,and gravity upon the acceleration of the object. An algorithm is presented toobtain an explicit expression for calculating these measures by decomposing thetorque-related non-square matrix into square matrices without using the conceptof pseudo-inverse for the planar multi-arm robot composed of arms each with twojoints.

Numerical examples for a planar robot with two 2R arms show thatthe proposed concepts are effective in representing the acceleration capabilityand velocity effects of a multi-arm robot.

One of the important characteristics of an autonomous robot lies in the capability of self-organizing its own behaviors in order to adapt to an unknown environment. The goal of this paper is to present an effective means for investigating such a capability, which enables us to graphically build a model of the autonomous robot and dynamically observe the emergence of reactive behaviors as a result of external stimulus controlled behavioral self-organization. In particular, the paper describes an integrated software workbench, called C4 (which stands for Create, Coordinate, Condition, and Co-evolve), for modeling and simulating autonomous robotic systems, e.g., legged and dual-arm ones. The module create enables us to build top-down graphical models of novel mechanisms, followed by primitive motion pattern (e.g., gaits) specification and visualization in the coordinate module. Learning mechanisms, as embedded in module condition further allow us to test how the animated robots acquire new behavioral rules triggerable by external stimuli. Finally, the module co-evolve facilitates the analysis of the distributed intelligence of robot groups, which manifests itself from the behaviors emergent from the interaction among individual robots.

This paper describes the experimental evaluation of three identification schemes to determine the dynamic parameters of a two degrees of freedom direct-drive robot. These schemes involve a recursive estimator while the regression models are formulated in continuous time. The fact that the total energy of robot manipulators can be represented as a linear relation in the inertial parameters, has motivated the suggestion in the literature of several regression models which are linear in a common dynamic parameter vector. Among them, in this paper we consider the schemes based on the filtered dynamic regression model, the supplied energy regression model and a new one proposed in this paper: the filtered power regression model. The underling recursive parameter estimator used in the experimental evaluation is the standard least-squares.

This paper identifies the need for a new range of end effectors suitablefor non-rigid products and introduces a novel non-contact gripping device. Theend effector operates on the principle of generating a high-speed flow betweenthe end effector and product surface thereby creating a vacuum which levitatesthe product. The lifting forces and conditions are theoretically discussed. Theexperimental results are presented using the end effector that has been operatedto lift non-rigid food materials such as jelly blocks.

In this paper, a systematic algorithm, the Parallel Scheme, is proposed to solve the multiple-goal problem of redundant manipulators. It can adaptively adjust all the weighting factors of the corresponding goals according to their individual needs. Therefore, conflicts due to redundancy competition are resolved. To upgrade the redundancy-release capability of the original Parallel Scheme, the Improved Parallel Scheme is also developed by using the exponential function to adjust the weightings. The Improved Parallel Scheme is shown to improve both the redundancy-release capability and the weighting-adjustment capability while retaining all the merits of the original Parallel Scheme.

During the past few years, there has been an increasing demand in thefield of precision engineering for fine motion in multi-degrees of freedomsystems. These applications motivated the development of a new robotics fieldcalled microrobotics. In this paper, we review both the design guidelines formicrorobots and the advantages of using parallel robots in very high precisionapplications. Parallel micromanipulators using elastic joints as well asstructures manufactured in single solid and metallic bellows areintroduced.