Ranging-Based Localizability Optimization for Mobile Robotic Networks

In robotic networks relying on noisy range measurements between agents for cooperative localization, the achievable positioning accuracy strongly depends network geometry. This motivates problem of planning robot trajectories in such multirobot systems a way that maintains high localization accuracy. We present potential-based methods, where localizability potentials are introduced to characterize quality geometry position estimation. These based Cramér Rao lower bounds (CRLB) and provide theoretical bound error covariance by any unbiased estimator. process, we establish connections CRLBs theory graph rigidity, which has been previously used plan motion networks. develop decentralized deployment algorithms appropriate large networks, use equality-constrained extend concept scenarios additional information about relative positions ranging sensors is known. illustrate resulting methodology through simulated examples an experiment.