Modular Manipulators for Cluttered Environments: A Task-Based Configuration Design Approach

Abstract

Modular manipulators gained popularity for their implicit feature of “reconfigurability”—that is, the ability to serve multiple applications by adopting different configurations. As reported in the literature, most of the robotic arms with modular architecture used specific values of twist angles, e.g., 0 deg or 90 deg. Further, the number of degrees-of-freedom (DoF) is also kept fixed. These constraints on the design parameters lead to a smaller solution space for the configuration synthesis problems and may result as no-feasible solution in a cluttered work-cell. To work in a realistic environment, the task-based customized design of a manipulator may need a larger solution space. This work deals with the extension of the modular architecture from conventional values to unconventional values of design parameters, keeping the degrees-of-freedom also as variable. This results into an effective utilization of modular designs for highly cluttered environments. A three-phase design strategy is proposed in the current work. The design strategy starts with the decision of optimal number of modules required for the given environment in the first phase, which is followed by task-based “configuration planning” and “optimal assembly” in the second and third phase, respectively. Three types of modules are proposed with same architecture and different sizes—heavy (H), medium (M), and light (L). The configuration planning includes detailed discussion on the type-selection of the modules and their possible combinations. Comparison of all possible n-link combinations is analyzed based upon the optimized results with respect to the minimum torque values. Case studies of a power plant with two different workspaces are included to illustrate the three-phase strategy representing the importance of modularity in nonrepetitive maintenance tasks.