Associated Sectors of Magnetic Recording Systems Using Spatially Coupled LDPC Codes

Abstract

In traditional magnetic recording systems, non-associated sectors are mainly adopted, whereby two consecutive sectors are decoded independently by the low-density parity-check (LDPC) codes. In this paper, we propose a magnetic recording system with associated sectors, constructed using spatially coupled low-density parity-check (SC-LDPC) codes. If the SC-LDPC decoder cannot correct the erroneous bits in the current sector, it can request information stored in previous sectors to improve decoding performance. Moreover, we modify protograph-based extrinsic information transfer (P-EXIT) charts to examine the theoretical performance of SC-LDPC codes applied to both non-associated and associated sectors. Our theoretical results show that the associated sectors achieve significant performance gains compared to the traditional non-associated sectors.

Sliding-Window Turbo Equalization and Its Reduced-Complexity Decoding Techniques

“…Previous work [19]- [24] used the C-TE, which replaces traditional LDPC codes with SC-LDPC codes. Fig.…”

“…3(a), the BCJR detector [25] , of each trellis section, involved in a surviving path, are produced, and passed to the SC-LDPC decoder. SC-LDPC decoding is performed using either the block method [19]- [22] or the sliding window method [23], [24]. Note that we consider the C-TE in this paper, in which the SC-LDPC decoder only uses window decoding [23], [24], as shown in Fig.…”

“…Over the last five years, SC-LDPC codes have emerged as an excellent outer code for ISI applications, such as magnetic recording systems [19]- [24]. In [19], Esfahanizadeh et al used SC-LDPC codes as the outer code in magnetic recording systems, and the Bahl-Cocke-Jelinek-Raviv (BCJR) [25] as the inner detector.…”

“…Moreover, dynamic window shifting was introduced to reduce the complexity of SC-LDPC decoding. Additional research of SC-LDPC codes with window decoding is considered in the turbo equalization of BPMR systems, as discussed in [24].…”

“…Based on the literature review, there is room to improve the performance of turbo equalization by modifying the message passing between the BCJR detector and SC-LDPC decoder. The first part of this paper is motivated by the fact that only the SC-LDPC decoder in turbo equalization uses window decoding, whereas the BCJR detector still produces messages with a complete block ensemble [23], [24]. In other words, the BCJR detector can only start processing when the SC-LDPC decoder has decoded all codewords, even though the SC-LDPC window decoder produces output nodes one by one.…”