This paper addresses the prediction of error floors of low-density parity-check codes transmitted over the additive white Gaussian noise channel. Using a linear state-space model to estimate the behavior of the sum-product algorithm (SPA) decoder in the vicinity of trapping sets (TSs), we study the performance of the SPA decoder in the log-likelihood ratio (LLR) domain as a function of the LLR saturation level. When applied to several widely studied codes, the model accurately predicts a significant decrease in the error floor as the saturation level is allowed to increase. For nonsaturating decoders, however, we find that the state-space model breaks down after a small number of iterations due to the strong correlation of LLR messages. We then revisit Richardson's importance-sampling methodology for estimating error floors due to TSs when those floors are too low for Monte Carlo simulation. We propose modifications that account for the behavior of a nonsaturating decoder and present the resulting error floor estimates for the Margulis code. These estimates are much lower, significantly steeper, and more sensitive to iteration count than those previously reported.

Recent work by Divsalarhas shown that properly designed protograph-based low-density parity-check codes typically have minimum (Hamming) distance linearly increasing with block length. This fact rests on ensemble arguments over all possible expansions of the base protograph. However, when implementation complexity is considered, the expansions are frequently selected from a smaller class of structured expansions. For example, protograph expansion by cyclically shifting connections generates a quasi-cyclic (QC) code. Other recent work by Smarandache and Vontobel has provided upper bounds on the minimum distance of QC codes. In this paper, we generalize these bounds to punctured QC codes and then show how to tighten these for certain classes of codes. We then evaluate these upper bounds for the family of protograph codes known as AR4JA codes that have been recommended for use in deep space communications in a standard established by the Consultative Committee for Space Data Systems. At block lengths larger than 4400 bits, these upper bounds fall well below the ensemble lower bounds.

Numerical issues related to the occurrence of error floors in floating-point simulations of belief propagation (BP) decoders are examined. Careful processing of messages corresponding to highly-certain bit values can sometimes reduce error floors by several orders of magnitude. Computational solutions for properly handling such messages are provided for the sum-product algorithm (SPA) and several variants.

Recent work has shown that properly designed protograph-based LDPC codes may have minimum distance linearly increasing with block length. This notion rests on ensemble arguments over all possible expansions of the base protograph. When implementation complexity is considered, the expansion is typically chosen to be quite orderly. For example, protograph expansion by cyclically shifting connections creates a quasi-cyclic (QC) code. Other recent work has provided upper bounds on the minimum distance of QC codes. In this paper, these bounds are expanded upon to cover puncturing and tightened in several specific cases. We then evaluate our upper bounds for the most prominent protograph code thus far, one proposed for deep-space usage in the CCSDS experimental standard, the code known as AR4JA.

This paper presents results from field tests of the CDMA system developed by Qualcomm, Incorporated. The CDMA system, now EIA/TIA digital standard IS-95, has undergone extensive field tests since its early inception. The latest series of technical tests was conducted in San Diego, California by Qualcomm in cooperation with several cellular carriers and equipment manufacturers during the month of August 1993. The goals of the tests were to provide an accurate measure of the capacity of the system, to validate the standard with compliant equipment, and finally to aid service providers in understanding the tradeoff of system parameters as large CDMA cellular networks begin to be deployed.>