CS-LSTM: a hybrid channel estimation scheme in IDFT-VFDM system

Recent inverse discrete Fourier transform-based Vandermonde-subspace frequency division multiplexing (IDFT-VFDM) is able to extend the uplink signal coverage for the popular new radio (NR) and future 6G wireless systems. Along this direction, channel estimation is the key for IDFT-VFDM to maintain its interference-free characteristics. However, it is challenging for traditional pilot-based channel estimation to obtain the perfect channel state information due to the limitation of IDFT-VFDM. To address this problem, we propose CS-LSTM, a novel channel estimation algorithm for IDFT-VFDM. The proposed CS-LSTM algorithm first estimates an instantaneous channel value through the preamble sequence, and then iteratively uses LSTM network and Compressed sensing algorithm to further estimate the channel value affected by subsequent Doppler frequency shifts. By doing so, our CS-LSTM is able to yield highly accurate state information. Simulation results show the effectiveness of the proposed CS-LSTM.