In this paper, we propose a new method to learn a multi-fingered hand grasping posture with little knowledge about the task and few sensing capabilities. The developed model is composed of two stages. The first is dedicated to the finger inverse kinematics learning in order to provide the fingertip-desired position. This function is fulfilled by modular neural network architecture. Following the concept of reinforcement learning, a second neural model dealing with noisy sensing information is used to search the space of hand configuration. Simulation results show a good learning of grasping postures with five fingers and different noise levels.

This paper demonstrates the convergence stability and the actual usefulness of the gradient-based motion optimizations for manipulator arms. An optimal motion-planning problem is converted into a finite-dimensional nonlinear programming problem that utilizes cubic or quintic B-splines as basis functions. This study shows that the numerically calculated gradient is a useful tool in finding minimum torque, minimum energy, minimum overload, and minimum time motions for manipulator arms in the presence of static or moving obstacles. A spatial 6-link manipulator is simulated without simplifying any of the kinematic, dynamic or geometric properties of the manipulator or obstacles.

In this paper, the optimum design of parallel kinematic toolheads is implemented using genetic algorithms with the consideration of the global stiffness and workspace volume of the toolheads. First, a complete kinetostatic model is developed which includes three types of compliance, namely, actuator compliance, leg bending compliance and leg axial compliance. Second, based on this model, two kinetostatic performance indices are introduced to provide a new means of measuring compliance over the workspace. These two kinetostatic performance indices are the mean value and the standard deviation of the trace of the generalized compliance matrix. The mean value represents the average compliance of the Parallel Kinematic Machines over the workspace, while the standard deviation indicates the compliance fluctuation relative to the mean value. Third, design optimization is implemented for global stiffness and working volume based on kinetostatic performance indices. Additionally, some compliance comparisons between Tripod toolhead and other two principal Tripod-based Parallel Kinematic Machines are conducted.

This paper explores a new type of a parallel platform human interface manipulator based on virtual reality (VR) for mechanism design applications. A motion control of a six-link robot manipulator actuated by three active joints is presented here. The main components of the system include a user interface, a software simulating the environment, and a VR control system. The model of the VR system is built based on a force feedback behavior that enables the operator to feel the actual force feedback from the virtual environment just as he/she would from the real environment. A primary stabilizing controller is used to develop a haptic interface device where realistic simulations of the dynamic interaction forces between a human operator and the simulated virtual object/mechanism is required. The stability and performance of the system are studied and analyzed based on the Nyquist stability criterion. Experiments on cutting virtual clay are used to validate the theoretical developments. It was shown that the experimental and theoretical results are in good agreement.

Fault diagnosis plays an important role in the operation of modern robotic systems. A number of researchers have proposed fault diagnosis architectures for robotic manipulators using the model-based analytical redundancy approach. One of the key issues in the design of such fault diagnosis schemes is the effect of modeling uncertainties on their performance. This paper investigates the problem of fault diagnosis in rigid-link robotic manipulators with modeling uncertainties. A learning architecture with sigmoidal neural networks is used to monitor the robotic system for off-nominal behavior due to faults. The robustness, sensitivity, missed detection and stability properties of the fault diagnosis scheme are rigorously established. Simulation examples are presented to illustrate the ability of the neural network based robust fault diagnosis scheme to detect and accommodate faults in a two-link robotic manipulator.

The paper presents research on the APF approach for solving the GNRON and local minima problems. The repulsive potential function is modified in order to solve the GNRON problem. A simulated annealing algorithm integrated into the APF has solved the local minimum problem. The improved APF is applied to the path-planning problem of soccer robots. The simulated experiments show the validity of this approach.

Sophisticated robotic applications require systems to be reconfigurable at the system level. Aiming at this requirement, this paper presents the design and implementation of a software architecture for a reconfigurable real-time multi-processing system for multi-robot control. The system is partitioned into loosely coupled function units and the data modules manipulated by the function units. Modularized and unified structures of the sub-controllers and controller processes are designed and constructed. All the controller processes run autonomously and intra-sub-controller information exchange is realized by shared data modules that serve as a data repository in the sub-controller. The dynamic data-management processes are responsible for data exchange among sub-controllers and across the computer network. Among sub-controllers there is no explicit temporal synchronization and the data dependencies are maintained by using datum-based synchronization. The hardware driver is constructed as a two-layered system to facilitate adaptation to various robotic hardware systems. A series of effective schemes for software fault detection, fault anticipation and fault termination are accomplished to improve run-time safety. The system is implemented cost-effectively on a QNX real-time operating system (RTOS) based system with a complete PC architecture, and experimentally validated successfully on an experimental dual-arm test-bed. The results indicate that the architectural design and implementation are well suited for advanced application tasks.

This paper describes precision enhancement of an optical three-axis tactile sensor capable of detecting both normal force and tangential force. The sensor's single cell consists of a columnar feeler and 2-by-2 conical feelers. We have derived equations to precisely estimate the three-axis force from the area-sum and area-difference of the conical feelers' contact areas by taking into account wrench-length shrinkage caused by a vertical force. To evaluate the equations and determine constants included in the equations, we performed a series of calibration experiments using a manipulator-mounted tactile sensor and a combined load-testing machine. Subsequently. to evaluate the tactile sensor's practicality. it was mounted on the end of a robotic manipulator which rubbed flat specimens such as brass plates with step-heights of δ=0.05, 0.1, 0.2 mm and a brass plate with no step-height. We showed from the experimental data that the optical three-axis tactile sensor can detect not only the step-heights but also the distribution of the coefficient of friction, and that the sensor can detect fine plate inclination with accuracy to about ±0.4°.

Ball passing is an elementary and frequently employed human soccer skill. This paper examines the realization and visualization of ball passing, a low level move-to-ball behavior of a soccer robot, in a robot soccer game. A case study of three mechanically identical mobile robots with a formation ready to pass a ball cyclically in a zigzag pattern is examined. We build a control command driven mobile robot motion simulator with a controller and dynamics of mobile robots, not only nonholonomic kinematic constraints to simulate the motion of a soccer robot driven by wheels torques to generate wheels accelerations, to update the robot position and orientation at successive time instants. Kick motion follows a physical law, and a simplified collision check and response model is utilized for the efficient detection of the hitting a robot with the ball or other robots. The realization of specific ball passing strategy to drive each soccer robot in a position to receive a pass includes three levels of organization, coordination, and execution: careful integrated design of a dynamic formation and role change scheme, ball position estimation, and coordinated trajectory (i.e. path and velocity) planning and tracking control. Simulations are performed to illustrate the feasibility of the realization of ball passing among three robots, implemented by a software program for coordinated trajectory planning and tracking control in the developed simulator.

Enhancing graphical objects whose behaviors are governed by the laws of physics is an important requirement in modeling virtual physical environments. In such environments, the user can interact with graphical objects and is able to either feel the simulated reaction forces through a physical computer interface such as a force feedback mouse or through such interactions, objects behave in a natural way. One of the key requirements for such interaction is determination of the type of contact between the user controlled object and the objects representing the environment. This paper presents an approach for reconstructing the contact configuration between two objects. This is accomplished through usage of the time history of the motion of the approaching objects for inverse trajectory mapping of polygonal representation. In the case of deformable objects and through usage of mass-spring-damper system this paper also presents a special global filter that can map the local deformation of an object to the adjacent vertices of polygonal mesh. In addition to offering a fast computational framework, the proposed method also offers more realistic representation of the deformation. The results of this paper are shown through detailed examples and comparison analysis using different computational platforms.

In this paper a hybrid parallel-serial manipulator, named as CaHyMan (Cassino Hybrid Manipulator), is analyzed in term of stiffness characteristics as a specific example of a general procedure for analyzing stiffness of parallel-serial manipulators. A formulation is presented to deduce the stiffness matrix as a function of the most important stiffness and design parameters of the mechanical design. A formulation is proposed for a stiffness performance index by using the obtained stiffness matrix. A numerical investigation has been carried out on the effects of design parameters and fundamental results are discussed in the paper.

The questions when and why one needs to use mathematical models, and especially the models of dynamics, represent still an unresolved issue. A general answer would be that dynamic modelling is needed as a tool when designing structure of the system and its control unit. In this case we talk about simulation. The other application is in on-line control of the system – the so-called dynamic control. While in simulation one generally uses the best available model, the control can be based on a reduced dynamics, depending on a particular task. The aim of this paper was to highlight the problems important for the dynamics of humanoids and their dynamic environment.

This paper presents the simplest walking model with an upper body. The model is a passive dynamic walker, i.e. it walks down a slope without motor input or control. The upper body is confined to the midway angle of the two legs. With this kinematic constraint, the model has only two degrees of freedom. The model achieves surprisingly successful walking results: it can handle disturbances of 8% of the initial conditions and it has a specific resistance of only 0.0725(−).

A novel affect-sensitive human-robot cooperative framework is presented in this paper. Peripheral physiological indices are measured through wearable biofeedback sensors to detect the affective state of the human. Affect recognition is performed through both quantitative and qualitative analyses. A subsumption control architecture sensitive to the affective state of the human is proposed for a mobile robot. Human-robot cooperation experiments are performed where the robot senses the affective state of the human and responds appropriately. The results presented here validate the proposed framework and demonstrate a new way of achieving implicit communication between a human and a robot.

This paper deals with an optimum synthesis of planar parallel manipulators using two constrained optimisation procedures based on the minimization of: (i) the overall deviation of the condition number of manipulator Jacobian matrix from the ideal/isotropic condition number, and (ii) bearing forces throughout the manipulator workspace for force balancing. A revolute jointed planar parallel manipulator is used as an example to demonstrate the methodology. The parameters describing the manipulator geometry are obtained from the first optimisation procedure, and subsequently, the mass distribution parameters of the manipulator are determined from the second optimisation procedure based on force balancing. Optimisation results indicate that the proposed optimisation approach is systematic, versatile and easy to implement for the optimum synthesis of the parallel manipulator and other kinematic chains. This work contributes to previously published work from the point of view of being a systematic approach to the optimum synthesis of parallel manipulators, which is currently lacking in the literature.

This article is proposing a grasp recognition and grasp mapping method for a land mine clearance master-slave system. The system consists of a data glove and a powerful hydraulic hand. Because of the different structure of the master and slave hand a mapping from master grasp to slave grasp has to be implemented. The paper presents a grasp recognition method by Neural Network (NN) and a mapping method. The introduced mapping method allows the operator more intuitive grasping.

This paper presents the control system of a chess-playing robot developed by the Research Institute of Robots at the Shanghai Jiaotong University. Thanks to the Windows NT operation system and the RTX (Real-Time eXtension), the whole system can achieve good real-time performance. The control system, which is supported by a standard PC hardware platform and a modularized structure of system software, is open-ended and easily expansible.

A new measurement device is proposed for the calibration of parallel manipulators that can be used to indentify all kinematic parameters with partial pose measurements. The device while restricting the motion of the end-effector to five degree-of-freedom measures three components of posture. A study is performed for a six degree-of-freedom fully parallel Hexa Slide Manipulator. Intrinsic inaccuracies of the measurement device are modeled with two additional identification parameters. Computer simulations show that all parameters, including the additional parameters, can be identified. Results show a significant error reduction, even with noisy measurements, and reveal that the identification is robust against errors in initial guess.

This paper describes emotion-based walking for a biped humanoid robot. In this paper, three emotions, such as happiness, sadness and anger are considered. These emotions are expressed by the walking styles of the biped humanoid robot that are preset by the parameterization of its whole body motion. To keep its balance during the emotional expressions, the motion of the trunk is employed which is calculated by the compensatory motion control based on the motions of the head, arms and legs. We have constructed a biped humanoid robot, WABIAB-RII (WAseda BIpedal humANoid robot-Revised II), to explore the issue of the emotional walking motion for a smooth and natural communication. WABIAN-RII has forty-three mechanical degrees of freedom and four passive degrees of freedom. Its height is about 1.84 m and its total weight is 127 kg. Using WABIAN-RII, three emotion expressions are experimented by the biped walking, including the body motion, and evaluated.

The experimental evidence supports the validity of the principle of superposition for multi-finger prehension in humans. Forces and moments of individual digits are defined by two independent commands: “Grasp the object stronger/weaker to prevent slipping” and “Maintain the rotational equilibrium of the object”. The effects of the two commands are summed up.