Direct ranging in multi-path channels using OFDM pilot signals

OFDM ranging is becoming important for positioning using terrestrial wireless networks. Conventional ranging methods rely on a two-step approach: range related parameters, such as the time of arrival (TOA), the bias induced by non-line-of-sight (NLOS) propagations etc., are first estimated, based on which the range is then inferred. In multi-path conditions, two-step range estimators which employ the correlator-based estimator or the energy detector lead to poor ranging accuracy when applied in non-ultra-wideband scenarios due to bias. More advanced ranging schemes that estimate all multi-path components using a multidimensional search procedure provide higher ranging accuracy but have a prohibitive complexity. In this work, we propose a novel direct ranging technique that uses a point process formulated channel model. Based on this model, we derive an approximate maximum likelihood estimator of the range. In contrast to the estimator which requires a multidimensional search procedure, the proposed estimator does not demand the knowledge of the exact number of multi-path components and these components are separable. If the power delay spectrum of the multi-path channel and the signal-to-noise-ratio (SNR) are known, the complexity of the proposed estimator is tractable. We show by means of Monte Carlo simulations that this estimator outperforms the correlator-based estimator.