Tsuneo Yoshikawa

Multifingered robot hands: Control for grasping and manipulation

Abstract Robot hands have been one of the major research topics since the beginning of robotics because grasping and manipulation of a variety of objects by robot hands are fundamental functionalities of various robotic systems. This paper presents a survey on the current state of research on control of grasping and manipulation by multifingered robot hands. After a brief history of the hardware development of multifingered robot hands, representative theoretical research results are presented in the area of grasping and manipulation. Regarding grasping, basic analytical concepts including force/form closures and active/passive closures are explained and various grasp quality measures for grasping position optimization are introduced. Regarding manipulation, the hybrid position/force control method and impedance control method are presented. Some of our recent results on grasping and manipulation by a soft-fingered hand are also presented. Finally, some future research directions are discussed.

A New End-effector for On-orbit Assembly of a Large Reflector

In Earth orbit, astronomical observations are possible free from any absorption or disturbances by the Earths atmosphere. Therefore, some large space telescopes and large space radio telescopes are planned for the future. We discuss the design of a radio telescope reflector which can be assembled in orbit by a space robot arm, with its networks and connecting mechanisms suitable for robot tasks. The characteristics of the new end-effector of robot arm and their suitability for onboard assembly tasks were confirmed by testing using a prototype end-effector and a two dimensional ground test arm. The test results are also described in this paper

Grasping Optimization Using a Required External Force Set

In this paper, we investigate optimal grasp points on an arbitrary-shaped grasped object using a required external force set. The required external force set is given based on a task, and consists of the external forces and moments, which must be balanced by virtue of contact forces applied by a robotic hand. When the origin is in the interior of the set, a force-closure grasp is required. When the dimension of the set is one, an equilibrium grasp is required. Therefore, we can investigate whatever the desired grasp is, such as when the desired grasp is a force closure and equilibrium grasps. Also, we only have to consider the forces contained in a given required external force set, not the whole set of possible resulting forces. Furthermore, we can avoid the frame-invariant problem (the criterion value changes with the change of the task (object) coordinate frame). We consider an optimization problem from the viewpoint of decreasing the magnitudes of the contact forces needed to balance any external force and moment contained in a given required external force set. In order to solve the problem, we present an algorithm based on a branch-and-bound method. We also present some numerical examples to show the validity of our approach. Note to Practitioners-This paper is concerned with grasping an object by a robotic hand. This article address how to grasp the object, namely, how to position every finger on the object. Recently, robots are desired to be used in housekeeping and in caring for elderly people. For this purpose, robot (multifingered) hands are equipped with the robots as general-purpose end effectors. The robot hands are required to automatically move to accomplish such tasks. In this case, the most fundamental issue for robot hands is to grasp the object. At home, there are many various-shaped objects. Consider the case where the robot (hand) is commanded to perform a certain task, such as putting the object into a box. In this case, the robot (hand) must grasp such an object (of any arbitrary shape) with appropriate grasp positions for completing the task. Therefore, the appropriate grasp positions must be calculated automatically. This article addresses a method to solve this problem. But to complete the grasping task, the following problems remain: calculation and control of the appropriate grasping forces

Controllability of Under-Actuated Planar Manipulators with One Unactuated Joint

This paper is concerned with analysis on controllability for a class of nonholonomic systems. We discuss controllability of underactuated planar manipulators with one unactuated joint. We show that if the first joint (in the base side) is actuated, these systems are completely controllable, namely, there exists an admissible trajectory from any given initial point to any given final point. In order to prove this, we use global stabilizing feedback control law to converge the state to a manifold, where the system is locally controllable. By this controller, we have two trajectories, one starting at the given initial position and the other starting at the given final position. Then we connect them using a kind of bi-directional approach to show the existence of the whole admissible trajectory. Finally, we give some simulation results to discuss controllability of more general cases, the first joint being actuated and all other joints being unactuated.

Shape recognition and grasping by robotic hands with soft fingers and omnidirectional camera

The purpose of this paper is to establish a method of shape recognition and grasping of unknown objects using a robotic hand with two soft fingers and one omnidirectional camera. For shape recognition, we propose to use a simple 2D version of the visual volume intersection method which takes advantage of a special feature of the omnidiretional camera attached to the hand. For grasping we propose a simple grasp quality criterion to obtain the best grasping position for the two soft fingers based on a visual hull in a horizontal grasp plane. We conducted several experiments and the results show the validity of the proposed method.

Optimization of power grasps for multiple objects

Power grasp is a grasp that can hold objects stably without changing the joint torques of fingers. Almost all studies on power grasp deal with one object, but it is more efficient to hold multiple objects at the same time. This paper derives a condition for power grasp for multiple objects, and defines an optimal power grasp from the viewpoint of decreasing the work of joint torques. Finally, we show some numerical examples to verify the validity of our approach.

Controllability of under-actuated planar manipulators with one unactuated joint

This paper is concerned with the analysis of controllability for a class of nonholonomic systems. We discuss the controllability of under-actuated planar manipulators with one unactuated joint. We show that these systems are completely controllable if the first joint (in the base side) is actuated. In order to prove this, we use a kind of bi-directional approach.

Path planning of a mobile robot for avoiding moving obstacles with improved velocity control by using the hydrodynamic potential

This paper describes the theory and the simulation of an improved velocity potential approach for path planning by which a mobile robot avoids standing and/or moving obstacles by using the hydrodynamic potential. This potential function for path planning is feasible for guiding a mobile robot to avoid an arbitrarily moving obstacle and to reach the goal in real time without finding the local maximum or minimum points in all cases. In this theory, there are two problems. One is that a mobile robot accelerates rapidly when it avoids a moving obstacle. The other is that a mobile robot has a discontinuous velocity when it is passing a moving obstacle. An ellipse field, which is obtained by using the conformal transformation, and a correction function, which generates the continuous velocity field, are installed in the previous potential function to cope with the difficulty. As a result, a mobile robot can gradually avoid a moving obstacle from further away, and can be safely guided without rapid acceleration.

Towards Whole Arm Manipulation by Contact State Transition

This paper discusses the whole arm manipulation allowing the contact state transition. For manipulation of an object under fully constrained, the contact state transition becomes necessary. In order to realize the object manipulation, we first derive the feasible direction of the object manipulation by analyzing the active/passive closure properties for every combination of contact states. Second, we derive the set of joint torque to move the object in the feasible direction. These analyses also provide the joint torque to realize the manipulation at the planned contact states. Effectiveness of the proposed method is confirmed by some simulation results

Shuffle turn and translation of humanoid robots

This paper proposes a novel shuffle-translating method that combines shuffle turns repeatedly for a humanoid robot. Conventionally, the walking motion of a humanoid robot is performed through a repeated foot stepping motion. However, this motion is inefficient and has low stability. Previously, we have studied the shuffle-turning method for a humanoid robot, which can perform a stepless and stable turning. In this paper, we present a precise shuffle turn with feedback control and a new shuffle translation method by repeating shuffle turns. Experiments using a humanoid robot were conducted and the results revealed that the proposed method was effective for shuffle translation.