Designing the social robot Elvis: how to select an optimal joint configuration for effective gesturing

Autor: Hoang-Long Cao, Greet Van de Perre, Bram Vanderborght, Albert De Beir
Přispěvatelé: Applied Mechanics, Robotics & Multibody Mechanics Research Group, Faculty of Engineering
Rok vydání: 2021
Předmět:
Zdroj: AIM
DOI: 10.1109/aim46487.2021.9517497
Popis: Depending on their exact application, social robots have been designed with different arm morphologies, ranging from under-actuated designs to arms featuring 9 degrees of freedom (DOF). It is however difficult to investigate if the chosen arm morphology is indeed the best possible solution for the intended application. Different design aspects, such as the number of DOF, the exact collocation of the joints, the link lengths and the joint angle range, all have a direct influence on the expressibility of the robot. Therefore, it is difficult to isolate the influence of one specific design parameter. To give insights in the effect of different design aspects on the performance of specified motions and help in making substantiated trade-off's in the design process of new robots, we developed a tool, based on the calculation of gestures for different morphologies and their visualization on one single virtual model. In this paper, we illustrate how this methodology was followed to select an optimal morphology for our new social robot Elvis. A preliminary gesture study indicated three joint configurations that could be particularly interesting for the intended application of our robot. Both the effect of individual joint placement, as well as the influence of joint angle range was investigated by performing a gesture study in simulation. To validate the output of the methodology, the three studied morphologies were physically realized and tested.
Databáze: OpenAIRE