Solving the dynamic equations of a 3-span style="text-decoration:underline; " class="text"P/spanRS Parallel Manipulator forefficient model-based designs

摘要

Introduction of parallel manipulator systems for different applications areashas influenced many researchers to develop techniques for obtaining accurateand computational efficient inverse dynamic models. Some subject areas makeuse of these models, such as, optimal design, parameter identification, modelbased control and even actuation redundancy approaches. In this context, byrevisiting some of the current computationally-efficient solutions forobtaining the inverse dynamic model of parallel manipulators, this papercompares three different methods for inverse dynamic modelling of a general,lower mobility, 3- style="text-decoration:underline; " class="text">PRS parallel manipulator. The first methodobtains the inverse dynamic model by describing the manipulator as three openkinematic chains. Then, vector-loop closure constraints are introduced forobtaining the relationship between the dynamics of the open kinematic chains(such as a serial robot) and the closed chains (such as a parallel robot).The second method exploits certain characteristics of parallel manipulatorssuch that the platform and the links are considered as independentsubsystems. The proposed third method is similar to the second method but ituses a different Jacobian matrix formulation in order to reduce computationalcomplexity. Analysis of these numerical formulations will provide fundamentalsoftware support for efficient model-based designs. In addition,computational cost reduction presented in this paper can also be an effectiveguideline for optimal design of this type of manipulator and for real-timeembedded control.