Utility of Aged Measurements for Localization

Many modern navigation scenarios involve autonomous agents navigating in an environment with a priori information of landmarks and obstacles. In such applications, agents make important navigation related decisions for avoiding obstacles. The agents often self-localize using Time-Of-Flight (TOF) measurements from known anchors by continuous sensor-polling. This polling uses system resources such as wireless spectrum and energy and a more opportunistic polling strategy would conserve these resources. For a previously obtained measurement, as the time elapses, the associated position information with this measurement ages. This is because the agent is continuously moving. Clearly, the nature of change of position information depends on the nature of the agent’s motion. In this work, we investigate the time dependence of position information as previously obtained measurements age. We discuss two general motion models - linear and circular, associated with the agent’s motion. Using the Cramer-Rao lower bound (CRLB), we analyze the effect of motion models and their influence on information content of aged sensor measurements. In particular for map based navigation systems, viewing an agent’s trajectory as a combination of linear and circular motion between waypoints, we analyze the variation of position error bound (PEB) with age. Finally, we present our insights about opportunistic sensor polling to a mab based navigation scenario.