Title:Towards Large-Scale Relative Localization in Multi-Robot Systems with Dynamic UWB Role Allocation

Abstract:Ultra-wideband (UWB) ranging has emerged as a key radio technology for robot positioning and relative localization in multi-robot systems. Multiple works are now advancing towards more scalable systems, but challenges still remain. This paper proposes a novel approach to relative localization in multi-robot systems where the roles of the UWB nodes are dynamically allocated between active nodes (using time-of-flight for ranging estimation to other active nodes) and passive nodes (using time-difference-of-arrival for estimating range differences with respect to pairs of active nodes). We adaptively update UWB roles based on the location of the robots with respect to the convex envelope defined by active nodes, and introducing constraints in the form of localization frequency and accuracy requirements. We demonstrate the applicability of the proposed approach and show that the localization errors remain comparable to fixed-role systems. Then, we show how the navigation of an autonomous drone is affected by the changes in the localization system, obtaining significantly better trajectory tracking accuracy than when relying in passive localization only. Our results pave the way for UWB-based localization in large-scale multi-robot deployments, for either relative positioning or for applications in GNSS-denied environments.