Abstract:

Due to extremely narrow pulses, an impulse radio signaling has a strong potential for high-precision positioning. Highly dispersive nature of ultra-wideband (UWB) channels makes time of arrival (TOA)estimation extremely challenging, where the leading-edge path is not necessarily the strongest path. Since the bandwidth of a received UWB signal is very large, the Nyquist rate sampling becomes impractical, hence motivating lower complexity and yet accurate ranging techniques at feasible sampling rates. In this paper, we consider TOA estimation baased on symbol rate samples that are obtained after a square-law device. Signal conditioning techniques based on wavelets and a bank of cascaded multi-scale energy collection filters are introduced, where correlations across multiple scales are exploited for edge and peak enhancements towards a more accurate detection. An adaptive threshold selection approach based on the minimum and maximum values of the energy samples is introduced, and optimal values of the thresholds for different signal to noise ratios (SNRs) are investigated via simulations. Theoreticaal closed form expressions are derived for mean absolute TOA estimation error, and compared with simulations. The performances of the discussed algorithms are tested on IEEE 802.15.4a residential line-of-sight (LOS) and non-LOS channels. Simulation results show that the introduced multi-scale energy product technique supported with a search-back step to detect the leading edge performs better than all the other techniques, excluding the pure threshold comparison algorithm at very large SNR values.