Abstract
Ultra-wideband (UWB) impulse radio is an attractive technology for ad hoc sensor networks due to enhanced interference mitigation, lower radiated power and accurate ranging capability. In the IEEE 802.15.4a standard, a time-multiplexed half-duplex ranging protocol between two distant nodes has been proposed to provide accurate ranging through time-of-flight measurements. To account for the crystal drift, the explicit ranging mode requires the receiving node to estimate the relative drift and to report it back to the initiator node. In this paper, we derive the maximum-likelihood (ML) drift estimator for an UWB pulse train signal. An estimation accuracy of 0.2-0.8 ppm can be achieved with a trellis-based ML design in severe multipath environments, where a sliding windowing scheme is implemented to keep the trellis memory to a minimum