Pneumatic actuator is low-weight, high-power and has backdrivability. In applications of the field of medicine, mechanisms using pneumatic actuators are studied for images acquisition. In images acquisition, ultrasound is non-invasive and enables to present the relative position between a therapeutic tool and organ etc. to physician in quasi-real-time. However, an ultrasound probe has individual specificity. Therefore, the relative position between the probe and an echogram was unclear and acquiring optional images of target object was difficult. Thus, we address the problem by 3-D image plane positioning method. In this study, we develop a probe scan mechanism for echography using pneumatic actuators and evaluate that basic performance of a position control accuracy of an image plane.
As the first step for the realization of automatic coupling of multiple vehicles, in this paper, we propose a simple approaching algorithm of two vehicles. The special feature of the proposed method is that the proposed method can bring a posteriorly-located crawler unit closer to the anteriorly located only utilizing low resolution sensors. Firstly, we derive an interesting geometrical relationship of two vehicles. This relationship correlate positions and attitude of each vehicles with Cross-deviation. Based on the derived geometrical relationship, we propose an approaching algorithm of two vehicles, Finally, using the linked crawler robot which is called DUCKS and a wireless communication module XBee, we confirm the feasibility of the proposed method.
The present paper proposes global quasi-static manipulation planning for the whole arm manipulation with a planar hand system and environments. The manipulation planning becomes complex since the changes of contact states between objects and hand/environments occur very frequently during the manipulation. In addition, a search space is expressed as a hybrid system which consists of time-continuous and discrete systems including motion states and contact states, respectively. To overcome those problems, a novel randomized algorithm is applied to the manipulation planning. The proposed planning generates a contact state network which connects the initial and goal configurations by expanding subgoals in the search space based on forward problem approaches. The validity of the proposed planning is shown by simulations.
In this paper we propose a dynamics error compensation method for the approximate gait in which the vertical ground reaction force varies such as a running gait. In this method, the dynamics error of the original gait is calculated as the moment of the ground reaction force about the target ZMP by using a more precise dynamics model. The allowable ranges of the horizontal ground reaction force and the ground reaction moment about the target ZMP is determined according to the vertical ground reaction force. The horizontal position and inclination of the torso are modified to reduce the dynamics error as well as to satisfy the allowable ranges using an inverted pendulum model and a flywheel model. The approximate gait is modified to have the total center of mass follow parabolic motion while conserving the angular momentum about the center of mass, by just setting both allowable ranges to [0,0] during the flight phase. The effect of the proposed method is shown by the simulation and experiment of ASIMO running.
This paper proposes a straightforward control raw for the position control of a two-links robotic arm in vertical plane with the gravitational force. This control strategy is not based on a conventional method using inverse kinematics and the Jacobian matrices of the robot. Virtually-desired joint trajectory is generated during the motion of the robot until the tip of the arm reaches a given target point. Because of the property, joint angle of the arm is not necessary to converge to the generated joint trajectory. Using this proposed control law, extremely smooth motion in position control can be realized like human arm movements, and resulting in Bell-shaped trajectory of velocity profile can be achieved through the PTP control. Usually, this kind of velocity profile has been seen in Minimum Jerk-Based Control and Minimum Torque-Based Control. However, the control law presented in this manuscript enables to produce the Bell-shaped velocity during the position control. Finally, this paper clearly indicates that an extremely simple control law acts as an effective method to mimic the natural and smooth motion of the human.