A Reference Nodes Selection Optimize Algorithm in Multi Robot System

Relative positioning is essential for control and coordination of multi-robot systems (MRS). For large-scale MRSs, relative positioning may consume too much computation and communication resources. By selecting a suitable subset of robots in the system as reference nodes (RN), the time difference of arrival (TDOA) algorithm can effectively resolve this problem. However, in a mobile MRS, varying topology may decrease the TDOA positioning accuracy. This paper addresses these challenges by presenting the dynamical RN selecting algorithm (DRS) in the fixed RNs amount situation. This algorithm updates the selection of RNs to maintain the accuracy of positioning when the topology changes. We prove the optimal conditions for RN selection with Cramer-Rao Lower Bound (CRLB) and propose an iterative RN selecting procedure. The simulations indicate that our algorithm can update the selection of RNs to limit the error of TDOA measurement efficiently.