Deep Residual Learning-Assisted Channel Estimation in Ambient Backscatter Communications

Channel estimation is a challenging problem for realizing efficient ambient backscatter communication (AmBC) systems. In this letter, channel in AmBC modeled as denoising and convolutional neural network-based deep residual learning denoiser (CRLD) developed to directly recover the coefficients from received noisy pilot signals. To simultaneously exploit spatial temporal features of signals, novel three-dimension (3D) block specifically designed facilitate CRLD. addition, we provide theoretical analysis characterize properties proposed Simulation results demonstrate that performance method approaches optimal minimum mean square error (MMSE) estimator with perfect statistical correlation matrix.