This paper discusses a path planning problem for a coordinated work by multi-robot system. The problem is how to move each robot's arm cooperatively to execute specified works. Various styles of coordinated work, which is based on the classification from a kinematic point of view for the relation of each robot's end-effector, are discussed. Mainly in this paper, a work while one robot is holding or carrying a workpiece, another robot executes a specified work with tools, is developed as a typical example of an interest coordinated work. The redundancy peculiar to the coordinated work is presented by the kinematic analysis of two-robot system. An algorithm to plan a coordinated motion is proposed by using the redundancy. As numerical examples, the proposed method is applied to a coordinated motion planning by two planar manipulators with three-links.
This paper describes a method for self-calibration of camera parameters of stereo vision systems.In stereo vision, for a point in one image its corresponding point lies on the so called epipolar line in the other image. The epipolar line is calculated from camera parameters. If the parameters include some errors, the errors cause small displacements between corresponding pairs and, thus, epipolar cons traint can not be used effectively. The errors of camera parameters can be calculated if a enough number of displacements of corresponding pairs are obtained. To detect enough displacements, an image is divided into small regions and at every region, a two-dimensional histogram of the disparity and the vertical displacement of features is generated, and the displacement is detected as the position of the prominent peak, if found. The detection of displacements is repeated until enough displacements are obtained everywhere in an image, where more reliable features are given a higher priority to detect reliable displacements earlier than less reliable ones. Since edges are used as features, structural descriptions are not necessary. Thus, this method is applicable to scenes with natural objects such as trees, bushes and so on. The calibration of displa cements provides establishing correct correspondences.
It is important that we control a robot manipulator considering contact with an object. In force control we have two difficulties caused by (1) environmental parameter change and (2) large contact force at collision. The collision always happens in a short transient period from trajectory control to force control. In this paper, we propose a robustifid force and collision control method. First, we designed TDOF (Two Degrees Of Freedom) force controller. This force controller enables us to obtain stable and fast response in the face of considerable environmental parameter variation. Next, in order to reduce the shock and also to change control modes smoothly from position to force, we proposed a novel collision control method by using feedforward position command together with the robust force controller based on position. We realized these proposals by using a TRANSPUTER based con-troller and confirmed their usefulness with experiments by using a 1-axis and a lifesize 6-axis mani-pulator.
This paper discusses an active tactile sensing not only for detecting a contact point between a robot and an unknown object, but also for detecting the object shape within the working area of the robot. We define the concept of Actively-Sensible meaning that an active sensing motion can be planned for a sensing event. We also define Minimum-External-Sensor-Realization for more than two sensing events, where a removal of any of pure external sensors always disables Actively-Sensible for any sensing events. When a robot hand approaches an unknown object and grasps it stably, it is necessary to know not only contact information but also surface information of the object. We explore object-shape-sensing as well as contact-point-sensing strategies which lead to the Minimum-External-Sensor-Realization, and eventually, propose a new algorithm for both sensings without using any pure external sensor. The proposed algorithm is verified by experiments using a newly designed single fingered robotic hand.
For motion planning of a manipurator, the use of a configuration space is highly efficient. Algorithms of the collision-free planning in static environments have been proposed in many papers. The efficiency of the algorithms depends on the description of obstacles in a configuration space. In general, the complexity of a configuration space map exponentially increases with the number of joints. Therefore, building the configuration space map is a time-consuming process, and the collision-free path can not be searched in real time. In this paper, we illustrate mathematical properties of a new kind of configuration space and propose an approach to parameterizing an obstacle in the configuration space. A spherical obstacle in a work space is transformed into a characteristic subspace in the configuration space. Therefore, we can approximate parameters of a sphere as a set of geometric entities. The transformation of a complex work space into the configuration space is described in terms of integrating the transformation of each element of a set of spheres. A collision-free path in a dynamic environment is found by using a Penalty-Function-Method. The Penalty-Function is modified according to the factors of the velocity along axes in a polar coordinates system.
It is more effective if a manipulator being set on a mobile robot can operate to load or unload while the mobile robot is moving. This paper presents the control method for the moving operations of an autonomous mobile manipulator (AMM) . The AMM moving in the horizontal plane produces position errors in the plane and rotation error around the vertical axis, but does not produce the errors concerning with the other directions. Using this relation, we present a hybrid control system which consists of a camera feedback system and a kinematical one. We designed a control system for the AMM by introducing switching matrix to avoid the interactions of the two control systems. The experimental results using our AMM which consists of a power wheeled steering mobile robot and a PUMA-type manipulator show basic data of the control system.
The major problem in binocular stereo vision is a well-known correspondence problem. Almost of the previous works solved it by regularization with the additional constraints such as the smoothness constraint or the ordering constraint. Therefore, they can not find the occlusion correctly and deal with the transparent surfaces, because these constraints are violated in the complicated scene contain-ing many occlusion. In this paper, we propose a method to detect binocular disparity and occlusion without using additional constraints by moving stereo cameras actively. It is based on the motion parallax obtained by the moving monocular camera. The search range of binocular disparity is restricted based on the monocular motion parallax. It uses only the Uniqueness of Disparity to find binocular disparity and occlusion. Therefore, it can find binocular disparity and occlusion correctly even for the complicated scene containing many occlusion. Furthermore, it can calculate binocular disparity very fast, because the calculation is only to count the number of disparity candidates in search range. Experimental results with complicated scene are presented to demonstrate the effectiveness of this method.
We have been developing a system which observes a human executing an assembly task, recognizes the task, and generates a robot program to achieve the same task. We call the system Assembly Plan from Obervation System (APO system) . In order to understand purpose of task operation, it is important to model a task execution procedure. This paper defines a task model for assembly of polyhedral objects. The task is analyzed based on face contacts between an object and other objects. The task model is defined as transition between face contact relations and operation which causes the transition. We also demonstrate an APO system based on the task model.
A pneumatic robot seems available to contacting tasks, because its compliant property caused by air compressibility is effective to absorb the collision force and to precisely control the contacting force. In this paper, the relationship between the inherent compliance due to air compressibility and the collision force is investigated. Also the compliance control method is discussed. The following is concluded. 1) To avoid the excessive collision force, joints are required to have as a low stiffness as possible. On this matter, a pneumatic robot is advantageous. 2) The proposed compliance control method being a hybrid of conventional position-based and force-based methods is useful to a contacting trace control. 3) A pneumatic robot can be expected to execute a stable contacting task even for fairly rugged object surface. This study shows the availability of a pneumatic robot to contacting tasks.
In this paper, it is shown that path planning of a robot manipulator whose workspace includes several obstacles can be done successfully by using a genetic algorithm. Several computer simulation results comfirm the effectiveness of our approach.