Maximizing safety margins in task-based design of redundant manipulators for cluttered environments

Kinematically redundant manipulators help in handling environmental constraints with extra degrees of freedom, but a large number of links may also lead to significant cumulative errors at the distal end, increasing the likelihood of collisions. The focus of this paper is to synthesize a robot with maximized tolerance to avoid potential collisions, while maneuvering in the workspace. A maximized-tolerance-based method in the design stage provides a significant margin to be utilized further during architectural planning and/or in error compensation against any joint clearance error. This is the main contribution of this paper. The strategy is applicable with even a large number of degrees of freedom. A measure, named as RoboGin, is defined both for a single configuration and for a set of configurations. Maximizing this metric over the large solution space of all robotic parameters provides an optimized design from the reliability perspective. The other requirements related to robot’s reachability at the specified task space locations (TSLs), kinematic conditioning and path connectivity are framed as constraints in the formulated optimization problem. The global solutions computed through a simulated annealing technique show significant improvements in overall safety margins even in highly cluttered environments and with a large number of links. Implementation of the proposed strategy is demonstrated through realistic cluttered environments of a power plant, for a leakage testing application.