A graph optimization approach to range-based relative location

Abstract With the development of multi-robot cluster system, research multi-node relative pose estimation is in ascendant. However, for environment without GNSS or fixed anchors, problem still challenging. In this paper, a general algorithm and system framework based on graph optimization proposed to fuse inter- intra-robot measurements, as well mobility models with sliding windows attitude estimation. order verify feasibility universality algorithm, has used design positioning UWB “single-node multi-tags” structure, which can simultaneously obtain reliable positions directions. The multi-source fusion further improve accuracy robustness positioning. Experiment results shows that harsh anchor points, reach centi-meter level, may be best potential choice multi robot complex environment.