Drive-train selection criteria for n-dof manipulators: Basis for modular serial robots library

Abstract

Towards planning a modular library for customized designs of serial manipulators, a trade-off is required between minimum modules inventory and maximum robotic applications to be handled. This paper focusses at the types of modules which are majorly based upon optimized payload capacity of the modular links. To find minimum types of modules in the modular library, an exercise has been performed on a large variety of robotic manipulators, with variations in degrees-of-freedom (dof) between 3 and 9 in number and that in payload capacity between 0 and 5 in kgs. Observing the pattern of the maximum-torque based drive-train selections for all the manipulators in consideration, three types of actuators are selected from a set of Maxon motor-gear assemblies. Subsequently, three types of modules are planned - Heavy (H), Medium (M) and Light (L). Challenge involved is the maximum load estimations for each joint involving variations due to large number of dof, various possible configurations and realistic weight estimation. This paper provides a general recursive framework for optimized drive-train, with one step as determination of maximum load estimation at a joint, and the second step as the selection of appropriate motor-gear assembly for the joint - providing an appropriate weight estimation for critical-configuration evaluation of the next link. The methodology is utilized for planning optimized number of modular divisions, for evaluating payload capacity of each division and possible modular combinations for given number of degrees-of-freedom.