Regarding assistance to disabled people for object manipulation and carrying, the paper focuses on the localisation for mobile robot autonomy. In order to respect strong low-cost constraints, the perception system of the mobile robot uses sensors of low metrological quality, ultrasonic ring and odometry. That poses new problems for localisation, in particular. Among different localisation techniques, we present only off-line localisation. With poor perception means, it is necessary to introduce a priori knowledge on sensors and environment models. To solve the localisation problem, the ultrasonic image is segmented applying the Hough transform, well-adapted to ultrasonic sensor characteristics. The segments are then matched with the room, modelled and assumed to be rectangular. Several positions are found. A first sort, based on a cost function, reduces the possibilities. The remaining ambiguities are removed by a neural network which plays the part of a classifier detecting the door in the environment. Improvements of the method are proposed to take into account obstacles and non-rectangular room. Experimental results show that the localisation operates even with one obstacle.

A fast numerical procedure is presented for computing growth distances between a pair of polytopes in three dimensional space that move incrementally along specified smooth paths. The procedure carrys out the growth distance evaluations efficiently by predicting and verifying contact configurations between a pair of grown polytopes. In the prediction and verification the procedure uses vertex and facial characterizations of polytopes and exploits their geometric adjacency information. The computation time, in average, is very small and does not depend significantly on the geometric complexity of two polytopes. A numerical example is presented to demonstrate the applicability of the procedure to interference detection in robotic simulations.

This paper introduces a simple method for obtaining the pure 3-D translational motion parameter (without rotation) and the translation parameter with a known rotation of a camera. Once the translation vector is available, the 3-D structure of an object can also be obtained. The minimum number of matching points in the two images can be as few as two, and the equations to be solved are linear. Synthetic and real image tests has been performed, and good results have been obtained.

This paper presents a novel approach to the hybrid position/force control of robot manipulators employed in executing repetitive tasks that involve contact with the environment. The constraint surface is assumed to be a plane and its respective dynamics is modelled by a mass-dashpot-spring scheme. Time results obtained by simulations are very encouraging for future research and developments, because the system's behaviour is very efficient.

GECCO-99 STUDENT WORKSHOP

(13th July, 1999)

The Genetic and Evolutionary Computation Conference (GECCO-99) in Orlando, Florida, USA was the biggest ever conference on evolutionary computation. It was a joint meeting of the long standing biannual International Conference on Genetic Algorithms (ICGA) which has been running since 1985 and the more recent annual Genetic Programming (GP) conference (which started in 1996). Before the start of each GP conference there has been a workshop for PhD students research in GP. This year Dr. O'Reilly (MIT AI lab) organized a joint graduate student workshop for research students principally concerned with any aspect of evolutionary computation, Evolvable Hardware, DNA computation, Artificial Life or Agents.

The design of the robotic mechanisms is most important because they determine the performance characteristics of the robots. This paper concerns the issue of computer-aided design (CAD) for planar 3-DOF parallel robotic mechanisms by means of the physical model of the solution space, which can be used systematically to express the relationships between the performance criteria and all link lengths of one type of the robtic mechanism. The performance atlases of the workspace volume for the manipulators are plotted in the physical model of the solution space. The characteristics of the distribution of the workspace shapes in the physical model of the solution space are presented. The results are useful for the optimum design of the robotic mechanisms. This paper proposes a new way for robotic CAD.

This paper presents a solution to the global adaptive partial state feedback control problem for rigid-link, flexible-joint (RLFJ) robots. The proposed tracking controller adapts for parametric uncertainty throughout the entire mechanical system while only requiring link and actuator position measurements. A nonlinear filter is employed to eliminate the need for link velocity measurements while a set of linear filters is utilized to eliminate the need for actuator velocity measurements. A backstepping control strategy is utilized to illustrate global asymptotic link position tracking. An output feedback controller that adapts for parametric uncertainty in the link dynamics of the robot manipulator is presented as an extension. Experimental results are provided as verification of the proposed controller.

This paper evaluates the influence of target location on robot repeatability. An experiment was set up to analyze the effect of the three-dimensional target location on robot repeatability. An error-analysis model to determine repeatability based on the robot's kinematic model and known robot parameters was developed. Experimental results indicated that there was a significant statistical difference between repeatability at different locations in the workspace and that the height of the target point influenced repeatability. Experimental results tended to those derived from the error-analysis kinematic model. Hence, to determine the optimal target location, there is no need for extensive experimentation; instead, only a few target points can be sampled and compared to an error-analysis model.

From a design point of view, it is crucial to predict singular configurations of a manipulator in terms of inputs in order to improve the dexterity and workspace of a manipulator. In this paper, we present a simple, yet a systematic appoach to obtain singularity contours for a class of five-bar planar parallel manipulators which are based on five rigid links and five single degree of freedom joints – revolute and prismatic joints. The determinants of the manipulator Jacobian matrices are evaluated in terms of joint inputs for a specified set of geometric parameters, and the contours of the determinants at 0.0 plane which are the singularity contours in joint space are generated for the three types of singularities reported in the literature. The proposed approach/algorithm is simple and systematic, and the resulting equations are easy to solve on a computer. The singularity contours for all the class are presented in order to demonstrate the method. It is concluded that the proposed method is useful in trajectory planning and design of five-bar planar parallel manipulators in order to improve their dexterity and workspace.