Representation of truss-style structures for autonomous climbing of biped pole-climbing robots

Abstract For autonomous pole-climbing robots, mapping and recognition of truss-style structures is a challenging task. To build an autonomous climbing system, a truss modeling and recognition system is proposed and applied to autonomous climbing. The system consists of three parts: environment modeling, segmentation by a proposed algorithm called Pouring Algorithm, as well as Truss Parametric Expression Algorithm (TPEA). The experiments show that our study is able to model the truss and extract the parametric expression with a 7.11 mm absolute error of the pole radius and a 9.38% relative error of the pole length. In addition, an autonomous climbing experiment based on our system with a climbing robot is illustrated, verifying the ability of our work to meet the requirement of autonomous climbing.