Best Viewpoints for External Robots or Sensors Assisting Other Robots

This work creates a model of the value different external viewpoints robot performing tasks. The current state practice is to use teleoperated assistant provide view task being performed by primary robot; however, choice ad hoc and does not always lead improved performance. research applies psychomotor approach develop relative quality using Gibsonian affordances. In this approach, for affordances are rated based on behavior human operators clustered into manifolds with equivalent value. 30 quantified in study 31 expert 4 (Reachability, Passability, Manipulability, Traversability) computer-based simulator two robots. adjacent similar values ranked agglomerative hierarchical clustering. results show validity affordance-based confirming that there statistically significantly viewpoint values, dependent affordances, independent robot. Furthermore, best manifold each affordance provides significant improvement large Cohen's d effect size (1.1-2.3) performance (improving time 14%-59% reducing errors 87%-100%) variation over worst manifold. will enable autonomous selection possible path planning