Abstract: For accurate time of arrival (TOA) estimations for the location based services (LBS) in the OFDM wireless networks, a novel TOA estimation algorithm for OFDM is presented. The TOA estimation is separated into the estimation of the integral sampling interval TOA part and the estimation of the fractional sampling interval TOA part. At the receiver, the characteristics of the training symbol in the time and frequency domain are exploited, and the fractional TOA part is obtained with iterative searches. This estimated fractional part is then adopted to mitigate the correlation spectrum spread evoked. Subsequently, the integral TOA part is achieved from this modified correlation spectrum by the leading edge method. Simulation results show that the novel algorithm outperforms the correlation method at SNRs of over 0dB. The estimation accuracy is at least twice that of the correlation estimation at the SNR of 10dB. Meanwhile, the super resolution TOA estimations with a complexity lower than one percent of the complexity, but with a comparable accuracy, of the available super resolution estimations are provided.

Key words: TOA estimation, Location Based Services, OFDM