Adaptive Time Synchronization between Transmitters in Digital Self-interference Cancellation Systems.

In 5G new radio, the number of band combination significantly increases and bandwidths are wider to meet a demand for high data rate. Hence, transmitter (TX) self-interferences significantly increase in a mobile front-end. Especially, in dual connectivity and carrier aggregation, intermodulation distortion product, which is generated by multiple TXs, directly lands in the receiver (RX) band and it severely degrades RX sensitivity. In full duplex technology, it may be more often because band combination becomes more diverse. Thus, TX self–interference mitigation is one of design issues in RF front-end due to an additional expense of high linearity components. Digital self-interference cancellation system has been developed to mitigate them. According to literatures, their performance is only guaranteed with time synchronization and it should be calibrated before the operation. But it is known that the time delay continuously varies. Time-varying delay between the reference interference model at TX and the interference in RX is usually compensated with the tapped delay line in the interference cancellation filter. But there is no efficient method yet to estimate time-varying relative delay between TXs. In this paper, an adaptive time synchronization method is proposed to estimate and compensate for time-varying relative delay between TXs using a least mean square approach. This proposed method has low-complexity because it jointly works with the existing interference cancellation filter and reuses its outputs. The degradation of signal-to-noise ratio (SNR) due to time-varying relative delay between TXs is more than 17dB at block error rate (BLER) 5%. Employing this method, SNR degradation at BLER 5% is improved to be within 0.2dB. And the tracking range of the relative time delay is guaranteed as long as the reference interference model is adequate to model TX self-interference.