TY - CHAP
T1 - Dynamic identification of a 3-prs parallel robot with scissor-type prismatic joint
T2 - A cloud-based approach
AU - Yang, Jian
AU - Wang, Xin
AU - Chen, Xue Feng
N1 - Publisher Copyright:
© Springer International Publishing Switzerland 2016.
PY - 2016
Y1 - 2016
N2 - 3-PRS parallel robot is now used for lots of applications. However, prismatic joints cannot satisfy some height limited situations. In order to decrease the vertical dimension of this type of robots, the scissor-type mechanism, which changes the vertical movement to horizontal, can be used to instead traditional prismatic pairs. Additionally, online identification of dynamic parameters is required when it comes to real-time control issues. Normally, common online identification approach, especially for parallel robots, lose accuracy for fast computation due to limitations of onboard computing and storage resources. In order to avoid those drawbacks of traditional online identification methods, a cloud-based approach, which leaves computing works to dynamic extendable resources with parallel computing capacities remotely, is suitable for this heavy computing task. Besides, cloud side storage can save all the data from the robots for machine-learning based algorithms or for remote fault diagnosis. In this paper, full dynamic model of this parallel robot is obtained by substituting scissor-types equivalent force to the Gibbs-Appell equation of the customary 3-PRS model. Then a cloud-based approach to online identification is introduced and the results appear in a good agreement between measured value and identified model outputs.
AB - 3-PRS parallel robot is now used for lots of applications. However, prismatic joints cannot satisfy some height limited situations. In order to decrease the vertical dimension of this type of robots, the scissor-type mechanism, which changes the vertical movement to horizontal, can be used to instead traditional prismatic pairs. Additionally, online identification of dynamic parameters is required when it comes to real-time control issues. Normally, common online identification approach, especially for parallel robots, lose accuracy for fast computation due to limitations of onboard computing and storage resources. In order to avoid those drawbacks of traditional online identification methods, a cloud-based approach, which leaves computing works to dynamic extendable resources with parallel computing capacities remotely, is suitable for this heavy computing task. Besides, cloud side storage can save all the data from the robots for machine-learning based algorithms or for remote fault diagnosis. In this paper, full dynamic model of this parallel robot is obtained by substituting scissor-types equivalent force to the Gibbs-Appell equation of the customary 3-PRS model. Then a cloud-based approach to online identification is introduced and the results appear in a good agreement between measured value and identified model outputs.
KW - 3-PRS parallel robot
KW - Cloud robotics
KW - Dynamic identification
UR - https://www.scopus.com/pages/publications/84948440297
U2 - 10.1007/978-3-319-23327-7_30
DO - 10.1007/978-3-319-23327-7_30
M3 - 章节
AN - SCOPUS:84948440297
T3 - Mechanisms and Machine Science
SP - 339
EP - 350
BT - Mechanisms and Machine Science
PB - Kluwer Academic Publishers
ER -