Cholesky Precoded Faster Than Nyquist Signalling with High Compression

Developing equalizers for Faster- Than-Nyquist (FTN) signals presents a significant challenge, and precoding offers a low-complexity alternative. In particular, employing a Cholesky factorization-based precoding enables us to achieve a symbol-by-symbol maximum likelihood decoding. Existing literature has established constraints on the FTN compression as a function of roll-off factors. This paper presents the feasibility of exceeding these conventional FTN compression limits while employing Cholesky precoding and ML decoding for FTN signals with small block sizes. Furthermore, it is shown that the bandwidth expansion is minimal in these precoded systems. Simulation results of the error performance of the precoded FTN system with Low-Density Parity Check (LDPC) codes are presented.