This paper describes a hierarchical architecture for rhythmic movement generation, which suits a juggling-like task involving sensory-motor coordination. Our approach, which is interpreted as a “bidirectional weak coupling” to the environment, does not require a continuous monitoring of the environment, but can adapt a robot to a change in the environment, owing to the interaction between the robot and the environment at the ball contact. The proposed architecture contains two passive-control mechanisms, the “entrainment mechanism” and the “open-loop stable mechanism,” that lead to the emergence of a self-organized temporal order in the whole system. This dynamic temporal pattern enables a robot to perform a stable rhythmic movement. We demonstrate a robot which juggles two balls rebounding off the wall and confirm the effectiveness of our approach.

Robot collision characteristics are analyzed by using the idea of the collision map. This analysis consists of the translations of the collision region on the collision map and they correspond to parallel movements of the original robot path. These translations are investigated in several cases and applied to general situations in which two robots are moving or working in a common workspace. Also, the collision characteristics are analyzed for a few special situations where the analysis of collision characteristics is crucial and the resultant solution for collision avoidance can be obtained.

This paper introduces a three degree of freedom XYZ Micromanipulator (XYZM) that is fabricated in the $x$-$y$ plane and positions components in the $x$, $y$, and $z$ directions using three independent linear inputs. The mechanism positions components on a platform using three legs, each composed of a slider mechanism and a parallelogram mechanism.

Three versions of the XYZM were fabricated and tested using surface micromachining processes: the rigid-body, offset, and compliant XYZM. Slider displacements of 45 micrometers result in a predicted out-of-plane displacement of 188 micrometers for the rigid-body XYZM, 205 micrometers for the offset XYZM, and 262 micrometers for the compliant XYZM.

A mathematic model is established to describe a swarm with multi-behavior. Regarding a swarm designed for cooperative task, we propose a model which includes a macroscopic model and a individual-based model. The macroscopic framework model describes global dynamics of swarms, which is normally expressed by dynamical populations' densities with different behaviors, while the individual-based framework model describes a individual agent's trajectory. Based on these models, we prove that all objects can be collected to the “home” area under conditions of individual agents subject to sensor constraints.

An efficient calibration-free visual servoing method is presented. The nonlinear optimization problem is formulated with the residual term for tracking a target. The residual term is estimated based on the secant approximation method. The image Jacobian is estimated using the affine model of the end-effector feature. Some experimental results are presented for a moving target.

The overall design of Ca.U.M.Ha. (Cassino-Underactuated-Multifinger-Hand) for harvesting horticulture products is presented. Ca.U.M.Ha. shows an anthropomorphic design incorporating four underactuated finger mechanisms and a simplified contrasting thumb, which are all joined to a rigid palm. The pneumatic cylinders of the articulated fingers are operated in parallel in order to give an additional auto-adaptability on the object to grasp. An application of Ca.U.M.Ha. for grasping different shapes of apples is presented.

A computational algorithm is developed to find a dynamic motion trajectory of a mobile manipulator with flexible links and joints that will allow the robot to carry a maximum load between two specified end positions. A compact form of the linearized state space dynamic equations is organized as well as constraint equations. Then, the problem of finding a maximum load carrying capacity on flexible mobile manipulators is formulated as a trajectory optimization problem.

This paper deals with the design and the experiments of the upper part of the ROBIAN biped robot. The motivation of the ROBIAN project is related to the study of the human being locomotion system. The major application of ROBIAN prototype is the development of a real testing bed of active/passive prosthesis devices enhancing research on the human being locomotion mechanism handicaps. The analysis of the wrench six components exerted by the upper part of a virtual manikin on the locomotion apparatus leads to the identification of two coupling relations. Based on the dynamic equivalence concept between mechanisms, the ROBIAN torso mechanism is identified. This concept leads to a four degrees of freedom mechanism able to reproduce the dynamic effects of the upper limbs during the walking gait. The mechanism parameters are optimized with respect to several design criteria and constraints. Then the prototype is built and mounted on the ROBIAN locomotion apparatus through a six components force sensor. Experimental results presented in this paper validate the proposed approach. The experimental coupling coefficients are identified. The influence of the masses motion on the ZMP trajectory are also given, showing the effectiveness of the torso mechanism contribution during a walking gait.

In this work, a hybrid adaptive/learning controller is simulated in Simulink using a two-link robot manipulator for a sinusoidal desired position. The two-dimensional nature of control law makes direct simulation difficult. Simulation results demonstrate that the controller achieves semi global asymptotic position tracking.

In this paper, the joint angles of a two link planar manipulator are calculated by using inverse kinematics equations together with some geometric equalities. For a given position of the end-effector the joint angle and angular velocity of the links are derived. The analyses contains many equations which have to be solved. However, the solutions are rather cumbersome and complicated, therefore a program is written in Fortran 90 in order to do, the whole calculation and data collection. The results are given at the end of this paper.

A two-degree of freedom control system that is most frequently encountered in practice is the so-called Internal Model Control (IMC) structure. However, the design procedure of such a structure does not present an easy task, which implies a limited utility of IMC. In this paper two alternative solutions are proposed that may be lumped together as Model-Following Control (MFC). These are two-loop control systems being easy to implement and offering interesting properties. Theoretical assumptions have been verified experimentally on a two-joint robot manipulator. Both qualitative and quantitative results yielded by experiments are presented and discussed.

A new fast successive approximation algorithm for the solution of the inverse position analysis of a general serial manipulator is presented. With the algorithm, we can search out the inverse solution of the serial manipulator quickly under the desired precision when the position of the three non-collinear end effector points is given. The position analysis of the 7R redundant serial manipulator is illustrated in the literature as an example. The simulation results verify the efficiency of the proposed algorithm. Since the three non-collinear end effector points can be selected at random, the algorithm can be applied to any other type serial manipulator.

The paper considers a technique for computation of the inverse kinematic model of the human arm for robot based rehabilitation that uses measurements of the hand position and orientation and radial acceleration of the upper arm. Analytical analysis and empirical validation of the method are presented. The algorithm enables estimation of human arm angles, which can be used in trajectory planning for rehabilitation robots, evaluation of motion of patients with movement disorders, and generation of virtual reality environments.

Two observers based on high order sliding mode approach are proposed to determine the absolute orientation of a walking biped robot without feet. Contrary to velocities observers which have been often designed for robot control, very few works have been proposed for the orientation estimation: in this paper, the estimation of all state variables are derived from only the actuated joint variables. Then the technology problem of the absolute measurement is avoided. This latter point is an original contribution of this paper. The observers and the control law converge in finite-time and are well adapted for analysis of the cyclic walking gait. Then, a second original contribution consists in adapting an existing “simplified” Poincaré's sections-based analysis of the stability of the walking to nonlinear systems with not fully available state variables.