Reduced-State Bahl–Cocke–Jalinek–Raviv Detector for Patterned Media Storage

Bit-patterned media (BPM) is a promising candidate for future high-capacity magnetic storage. The inter-track interference (ITI) in BPM has been taken into account by the “multiple islands per read head” model. However, the complexity of the symbol-based detection for this model is too high for practical use. To reduce complexity, we propose to use the reduced-state Bahl-Cocke-Jelinek-Raviv (BCJR) detection algorithm for such a two-dimensional model. We compare the computational complexity of the proposed reduced-state detector to that of the full-state BCJR detector. Simulation results show that the proposed reduced-state BCJR detector can approach the optimal performance by 0.5 dB while significantly reducing the detector complexity.