Asynchronous Ambient Backscatter Communication Systems: Joint Timing Offset and Channel Estimation

Ambient backscatter communication (AmBC) has attracted attention as an enabling technology for green device-to-device (D2D) communication for the next generation of Internet-of-Things (IoT) networks. Most existing works assume that the signals of the source and the backscattering device (BD) are perfectly synchronized at the reader. Perfect synchronization is not feasible in practice, and the timing offset between the two signals can significantly degrade the performance of the reader's receiver. In this work, we consider an asynchronous AmBC system and solve the problem of joint timing offset and channel estimation at the reader. Assuming generic Nyquist pulse-shaping filters, we develop a pilot-based maximum likelihood (ML) estimator, as well as a semi-blind estimator using the expectation maximization (EM) algorithm. For the special case of the rectangular pulse, we also derive the corresponding ML and EM estimators, and an approximate low-complexity EM (LCEM) algorithm. As theoretical benchmarks, we obtain the Cramer-Rao-bound (CRB) for pilot-based estimation, while for semi-blind estimation the modified CRB (MCRB) is obtained. The performance of the proposed algorithms is investigated using simulations, showing that both the ML and EM approach their corresponding CRBs, and that the EM-type algorithms provide superior estimation and detection accuracy, at the expense of higher complexity.