Time-varying Channel Modelling and Estimating for Reliable Edge and Cloud Computing Networks

Edge and cloud computing networks requires reliable transmission links, but facing the challenges of highly dynamic channels if high mobility nodes are present. The OTFS (Orthogonal Time Frequency Space) modulation has been recognized as a promising transmission technique to establish reliable communications. The underlying delay-Doppler (DD) domain channel used in OTFS is critical for the application. However, the relationship between the DD domain channel and the traditional time-delay channel is not clear, and the efficient channel estimation method is demanded. In this paper, we start from the time-delay multipath channel, and show that the traditional LTI model is just a zero-order Taylor expansion of the time-varying channel. In addition, we demonstrate that the DD domain channel model is also an approximation of the time-varying channel based on the first-order Taylor expansion by ignoring the scaling effect. By revealing the relationship between the time-frequency domain channel and the DD domain channel, we proposed an efficient channel estimation method to exploiting the DD domain sparsity with sympletic finite Fourier transform (SFFT). Numerical results show the superiority of the DD domain model in time-varying channels and the proposed channel estimation method over the traditional OFDM framework.