학술논문

Assist Effectiveness Study based on Viscosity: Comparison of Assumed Command Signal and Actual Command Signal
Document Type
Conference
Source
2022 31st IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) Robot and Human Interactive Communication (RO-MAN), 2022 31st IEEE International Conference on. :686-692 Aug, 2022
Subject
Communication, Networking and Broadcast Technologies
Computing and Processing
Robotics and Control Systems
Signal Processing and Analysis
Viscosity
Legged locomotion
Torque
Switches
Muscles
Stairs
Angular velocity
Language
ISSN
1944-9437
Abstract
In wearable assist devices, the assumed and actual command timings are different. This discrepancy cause problems, such as a decrease in the assist effect. In this study, we focus on viscosity, which is a characteristic of human muscles, and propose viscosity assistance to solve the problems. This method outputs the torque based on the predetermined viscosity coefficient and the actual angular velocity if the torque and angular velocity of the human joints are in opposite directions. The possibility of viscosity assist is demonstrated by analyzing the seated movement, and an assist device using a magnetorheological fluid brake is developed. The assist device is driven using three command inputs: time-based, joint angle-based, and constant viscosity coefficient command inputs. Under conditions based on time and angle, the actual output deviates from the assumed value when the motion is different from the expected motion. In terms of the viscosity command, when only the constant command is used throughout the entire motion section, some subjects showed almost identical results to those expected. Even if a deviation from this assumption is indicated, the degree of agreement can be improved by switching several types of viscosity coefficients during motion.