A human-demonstration based approach to the recognition of process state transitions in the insertion of a deformable tube is presented. First, human demonstration of the insertion of a deformable hose into a rigid plug is measured by use of a position sensor and a force sensor. Secondly, the measurements are analyzed with regard to the process states during the insertion. Human recognition laws to detect process state transitions from force sensation are then extracted through the analysis of human demonstration. Finally, human motion during the insertion is transplanted to a mechanical manipulator to examine whether the human recognition laws are useful or not for the insertion performed by mechanical manipulators.

A static analysis of deformable object grasping based on bounded force closure is presented. There are many manipulative operations that deal with deformable objects in manufacturing processes. Manipulative operations for these objects are often performed by utilizing their deformation actively while the operations may result in failure because of unexpected deformation of the objects during the manipulation process. In order to perform the manipulative operations for deformable objects successfully, it is necessary to evaluate their deformation by building object models and to derive task strategies by analyzing manipulation processes using the object models. In this paper, we will analyze stable grasping of deformable objects based on the concept of bounded force closure. Firstly, we will introduce the concept of bounded force closure, which is an extension of force closure condition. Secondly, we will investigate the necessary condition for bounded force closure in order to derive the properties of bounded force closure grasping. Thirdly, we will formulate the deformation of linear objects as an example of deformable objects and we will propose a procedure to evaluate stability of deformable object grasping. Finally, some numerical examples will be shown in order to demonstrate the effectiveness of our proposed method.

One approach to the learning of admittance matrix elements in, damping control is presented. It is difficult to determine admittance matrix elements analytically since they depend upon many physical properties, which are not identified easily, of manipulated objects. We propose a method to obtain admittance matrix elements through iterative trials of a manipulative operation. First, mating processes are kinematically analyzed and conditions for admittance matrix elements are derived based on modifiability analysis. Second, reaction forces during a mating process are investigated and a function to evaluate the process is defined. A learning algorithm, is then, developed based on nonlinear programming techniques. The developed algorithm is applied to a mating operation in a computer simulation. Finally, an experimental result is shown to demonstrate the validity of the proposed method.

One approach to the transplantation of human skilful motion to a manipulator is presented. First, human motion during the insertion of a deformable hose into a plug is measured by use of a position sensor and a force sensor. The measurements are then analyzed with regard to the contact between the hose and the plug. It is found that an alternating motion during the process is effective on the successful insertion of a deformable hose into a plug. Next, human motion is embedded in a manipulator program directly. It turns out that the manipulator can perform the insertion successfully as far as the positional error of the hose is involved in a certain range. Finally, it is shown that refining human motion helps a manipulator perform the operation more quickly.

Various deformable objects are manipulated in many manufacturing processes. Deformation of these objects is often utilized in order to manipulate them successfully while the manipulation sometimes fails because of unexpected deformation of the objects. Modeling of deformable objects is thus required so that the shape of the objects can be evaluated on a computer in advance. In this paper, we develop an analytical method to model the shape of a deformable linear object such as cords and tubes. First, a geometric representation to describe the shape of a linear object with bending and torsional deformation is introduced. The potential energy of the object and the geometric constraints imposed on it are then formulated. The shape of the object in the stable state can be derived by minimizing the potential energy under the geometric constraints. Next, procedure to compute the deformed shape is developed by applying a nonlinear programming technique. Finally, some numerical examples are shown in order to demonstrate how deformed shapes of linear objects are computed using the proposed approach.

Various deformable parts such as cords, leather products, and sheet metals are manipulated and are handled in many manufacturing processes. Deformation of these parts is often utilized in order to manipulate them successfully while the manipulation sometimes fails because of unexpected deformation of the parts. Modeling of deformable objects is thus required so that the shape of the soft parts can be analyzed and evaluated on a computer. In this paper, we develop an analytical method to model the shape of a deformable object. Especially, we deal with deformation of a bendable thin object. The process of manipulating a deformable object is analyzed with regard to how the object interacts with other objects around it. The model of a bendable thin object is formulated according to the principle that the potential energy of the object reaches the minimum at its stable shape. An algorithm to compute the deformed shape of the object is developed by applying a nonlinear programming technique. Finally, a simple experiment is done to demonstrate the validity of the modeling method proposed in this paper.>

The estimation of contact states by using force information acquired in the mating process is discussed, and a method for generating the state classifiers based on geometric models of workpieces on a computer is developed. A symbolic representation of contact states is addressed. Static behavior of workpieces at each contact state is analyzed by applying the theory of polyhedral convex cones. State classifiers that discriminate contact states are formulated by using the polyhedral convex cones, which directly provide a set of discriminant functions. To reduce real-time computations, the classifiers are simplified to a minimum set by using reduction rules of polyhedral convex cones. The algorithm to generate the state classifiers was implemented on a computer. An experiment to identify the current contact state from the measured reaction force is described to demonstrate the usefulness of the approach.>

A model-based approach to the generation of damping control law for part-mating is presented. Partmating operation must be achieved despite positioning errors of workpieces. Applying d amping control, a robot can m odify its motion according to reaction forces acting on it in order to achieve the operations. A modification matrix, which characterize the d amping control law, depends upon t he operation to be done. Few research has been, however, studied about the derivation of modification matrices suitable for operations. Thus, a method to derive modification matrices from the geometric model of workpieces is developed. First, the process of part-mating operation is modeled as a set of transitions of contact states. Second, static property at each contact state is analyzed. Desired motion set is then formulated in order to derive conditions for modification matrices that reduce the positioning error between workpieces. These conditions can be solved by applying the theory of polyhedral convex cones.

The human body motions during standing up and sitting down from/to a chair are analyzed employing a minicomputer system equipped with a graphic tablet and CRT displays. A two-dimensional mathematical model of the human body is developed and its equa-tions of motions are solved to evaluate forces and torques acting at each body joint. The computer graphics are successfully utilized throughout the study in an interactive manner. The effect of aging upon the body motion is also discussed.