Feasibility of Single-Agent Localization from Sequential Measurements

Localization from a sequence of measurements is feasible when the localization error does not increase over time. Under the widely adopted Bayesian filtering perspective, the computation of this bound and the proof of its underlying results become intricate. This paper characterizes the necessary conditions for feasible localization of a single agent. In particular, we find closed- form expressions for the necessary conditions that the position or distance measurements must satisfy in localization systems to guarantee that the error bound does not increase over time. We corroborate the obtained expressions with numerical results revealing the usefulness of the conducted research in a case study of wireless localization.