This paper proposes a passivity-based nonlinear disturbance observer (DOB) design for a powered upper-limb robot control. The proposed DOB allows for the nonlinearities of the robot dynamics, whereas the typical DOB designs cannot. Moreover, by virtue of the passivity property, human operator and environmental interactions can be embedded in the control loop. As a DOB, the proposed approach has a disturbance observation property that makes the actual robot behave like a nominal model selected by the user. Performance analysis proposes a gain tuning rule. In experimental validation, actual powered upper-limb robot is used to perform unknown payload carrying applications.

In impact dynamics, coefficient of restitution (COR) is an important parameter to calculate the impulse. Generally, COR has been assumed a constant value. However, COR varies according to the colliding condition. In order to apprehend this phenomenon, analysis of impact dynamics is conducted using a 3-DOF finger mechanism. The external impulse exerted on a wall or a cart by the mechanism is a function of the robot's geometry and dynamic parameters. In this paper, we derive an analytical model of COR with the 3DOF finger model colliding to a wall and a cart, and emphasize that the actual motion just after impact is not the same as what we expected unless the variable COR is considered. The effectiveness of proposed COR model is verified through wall and cart impact experiments using the finger mechanism.

This paper deals with the impact dynamics of a soccer ball. The trajectory of the ball is decided according to its initial velocity, launch angle, and some aerodynamic effect. Specially, the initial velocity is created by an instant impulse given to the ball. The external impulse exerted on a ball by a kicking robot is a function of the robot's geometry and dynamic parameters. Here, the coefficient of restitution (COR) is an important parameter to calculate the impulse applied to the ball when the collision is occurred. Usually, COR has been assumed a constant value. We derive an analytical model of COR for the pendulum model colliding with a soccer ball, and show that COR varies according to the velocity. The drag force is taken into account in the COR model. The velocity-dependent characteristic of COR is applied to simulation of a soccer ball traveling a long distance. It is shown that the velocity-dependent COR considerably affects the traveling distance of the flying ball.

In tightly coupled cooperative manipulation of two arms, an interaction force control is one of the important issues. In this paper, a sensor-less interaction force control in a de-burring task using a dual-arm is introduced. The dual arm manipulation increases the motion dexterity in the de-burring task as compared to the single arm manipulation. A closed form interaction force control algorithm is proposed, which does not require any force sensor. In the motion planning, the 12 Lagrangian coordinates of the system are decomposed into 9 motion degrees and 3 constrained degrees. Then, the whole dual arm can be modeled as a kinematically redundant robot. The primary task is to control the 9 motion degrees and the secondary task is to control the interaction force and the relative motion between the two arms. The concept of virtual joints is employed to represent the constrained degrees, which include the interaction force and the relative motion between the two arms. Through the simulation results, we show that the smooth de-burring task can be accomplished by using the suggested interaction force model without force sensor.