Time-Delay Estimation for Self-Interference Cancellation in LTE-A/5G Transceivers

Transmitter-to-receiver (Tx-Rx) leakage is a widely covered downside of state-of-the-art frequency division duplex radio frequency transceivers for use in mobile communication devices. Despite the distance between Tx and Rx carrier frequencies, non-idealities of the analog front-end cause receiver desensitization by folding transmit signal components into the Rx baseband. In literature, several countermeasures for this self-interference issue have been proposed, including fully-digital and mixed-signal mitigation strategies. Both methods employ signal estimation techniques in the digital domain to replicate and cancel the interference. An apparent issue of these methods is the unknown and usually time-varying delay of the leakage signal through the analog front-end of the device. Insufficient alignment of the signals used by the estimator causes severe degradation of the cancellation performance. In this work, we provide a mathematical analysis of a linear system identification scenario in the presence of an alignment mismatch. Based on these results, we present two low-complexity algorithms for static time-delay estimation and online tracking, accompanied by suitable digital hardware implementations. With focus on the particularly challenging signal statistics of Long Term Evolution (LTE) signals, we show the expectable performance of the algorithms for a specific linear self-interference cancellation task.