No abstract

This paper proposes a method of detecting the postural stability of a person wearing the lower limb exoskeletal robot with the HAT(Head-Arm-Trunk) model. Previous studies have shown that the human posture is stable when the CoM(Center of Mass) of the human body is placed on the BoS(Base of Support). In the case of the lower limb exoskeletal robot, the motion data, which are used for the CoM estimation, are acquired by sensors in the robot. The upper body, however, does not have sensors in each segment so that it may cause the error of the CoM estimation. In this paper, the HAT(Head-Arm-Trunk) model which combines head, arms, and torso into a single segment is considered because the motion of head and arms are unknown due to the lack of sensors. To verify the feasibility of HAT model, the reflecting markers are attached to each segment of the whole human body and the exact motion data are acquired by the VICON to compare the COM of the full body model and HAT model. The difference between the CoM with full body and that with HAT model is within 20mm for the various motions of head and arms. Based on the HAT model, the XCoM(Extrapolated Center of Mass) which includes the velocity of the CoM is used for prediction of the postural stability. The experiment of making unstable posture shows that the XCoM of the whole body based on the HAT model is feasible to detect the instance of postural instability earlier than the CoM by 20-250 msec. This result may be used for the lower limb exoskeletal robot to prepare for any action to prevent the falling down.

No abstract

The fluid gear pump-based system has repetitive disturbances, such as flow ripples, due to the mechanical characteristics of the gear system. The periodic disturbances have negative effects on the precise flowrate control, which is essential for consistent coating quality in the painting process. This study proposes a precise flowrate control method of the fluid gear pump-based painting system to compensate for the periodic disturbances. The compensation value of the controller output can be obtained by a repetitive controller. A compensation lookup table corresponding to the reference speed and the rotation angle can be generated through the repetitive controller. In order to secure robustness against various situations, a closed-loop system consists of the conventional proportional-derivative (PD) controller and a compensation lookup table in the form of the feedforward controller. The lookup table-based feedforward controller was compared with the open-loop controller and PD controller. Experimental results show that the proposed method is more effective than existing controllers in terms of periodic disturbance compensation. By using the results of this study, it is possible to improve the performance of the fluid gear pump-based painting system and precisely control the paint spray amount.

No abstract