Abstract
The fifty-year progress of faster-than-Nyquist (FTN) signaling is surveyed. FTN signaling exploits non-orthogonal dense symbol packing in the time domain for the sake of increasing the data rate attained. After reviewing the system models of both the conventional Nyquist-based and FTN signaling transceivers, we survey the evolution of FTN techniques, including their low-complexity detection and channel estimation. Furthermore, in addition to the classic FTN signaling philosophy, we introduce the recent frequency-domain filtering and precoding aided schemes. When relying on precoding, the information rate of FTN signaling becomes related to the eigenvalues of an FTN-specific intersymbol interference matrix, which provides a unified framework for the associated information-theoretic analysis and simplifies the associated power allocation specifically designed for increasing the information rate attained. We show that the FTN signaling scheme combined with bespoke power allocation employing a realistic raised-cosine shaping filter achieves the Shannon capacity associated with ideal rectangular shaping filters.
Highlights
Since the 1960s [1]–[3], faster-than-Nyquist (FTN) signaling has attracted researchers in the field of communications, owing to its potential of increasing the information rate beyond the bound defined by the classic Nyquist criterion
Motivated by the classic waterfilling algorithm developed for an singular value decomposition (SVD)-precoded spatial multiplexing multiple-input multiple-output (MIMO) system, in [15] both the optimal and a truncated power allocation schemes were proposed for eigenvalue decomposition (EVD)-precoded FTN signaling, in order to increase the achievable information rate
The first information-theoretic analysis of FTN signaling was presented by Rusek and Anderson in [5], where the capacity of unprecoded FTN signaling in an additive white Gaussian noise (AWGN) channel was derived by extending the classic capacity of Nyquist signaling over an inter-symbol interference (ISI) channel, as follows [5]: 1/(2τ T0)
Summary
Since the 1960s [1]–[3], faster-than-Nyquist (FTN) signaling has attracted researchers in the field of communications, owing to its potential of increasing the information rate beyond the bound defined by the classic Nyquist criterion. With the aid of the recent eigendecomposition-based TPC [14], [15], [18], [44], it becomes possible to consider an ISI-free received signal model having parallel substreams This allows us to formulate the information-theoretic analysis of FTN signaling in a simplified manner and obtain novel additional insights beyond those emerging from the early FTN studies. The use of the rectangular filter of (17) allows the ISI-free maximum symbol rate in a bandlimited channel, it is a challenging task to implement an ideal sinc pulse [4], [56] To circumvent this limitation, typically a raised-cosine (RC) shaping filter is employed [55], [56]. FTN SIGNALING In Section II, we reviewed the classic Nyquist criterion, which has been typically employed in conventional communication systems for reliable ISI-free information transmission. We introduce the general system model of FTN signaling and review the early studies
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have