We propose a maximum toroidal distance (MTD) code for lattice-based public-key encryption (PKE). By formulating the encryption encoding problem as the selection of $2^\ell$ points in the discrete $\ell$-dimensional torus $\mathbb{Z}_q^\ell$, the proposed construction maximizes the minimum $L_2$-norm toroidal distance to reduce the decryption failure rate (DFR) in post-quantum schemes such as the NIST ML-KEM (Crystals-Kyber). For $\ell = 2$, we show that the MTD code is essentially a variant of the Minal code recently introduced at IACR CHES 2025. For $\ell = 4$, we present a construction based on the $D_4$ lattice that achieves the largest known toroidal distance, while for $\ell = 8$, the MTD code corresponds to $2E_8$ lattice points in $\mathbb{Z}_4^8$. Numerical evaluations under the Kyber setting show that the proposed codes outperform both Minal and maximum Lee-distance ($L_1$-norm) codes in DFR for $\ell > 2$, while matching Minal code performance for $\ell = 2$.

Module Learning with Errors (M-LWE) based key reconciliation mechanisms (KRM) can be viewed as quantizing an M-LWE sample according to a lattice codebook. This paper describes a generic M-LWE-based KRM framework, valid for any dimensional lattices and any modulus$q$without a dither. Our main result is an explicit upper bound on the decryption failure rate (DFR) of M-LWE-based KRM. This bound allows us to construct optimal lattice quantizers to reduce the DFR and communication cost simultaneously. Moreover, we present a KRM scheme using the same security parameters$(q, k, \eta_{1}, \eta_{2})$as in Kyber. Compared with Kyber, the communication cost is reduced by up to 36.47% and the DFR is reduced by a factor of up to 299. The security arguments remain the same as Kyber.

This paper presents a majority-logic decoding (MLgD) algorithm for non-binary LDPC codes based on a novel expansion of the Tanner graph. The expansion introduced converts the Q-ary graph into a binary one, which makes the new MLgD algorithm more attractive for hardware implementations. Proposed algorithm performs significantly better than the existing MLgD algorithms in the waterfall region, and it shows a much lower error-floor as well. Algorithm only requires integer additions, comparisons, finite field operations and some binary operations. Thus, it offers an effective trade-off between performance and complexity in decoding non-binary LDPC codes.

In multiple-input multiple-output (MIMO) channels with discrete input alphabets, at high signal-to-noise ratio (SNR), maximizing the minimum Euclidean distance (dmin) between all possible received constellation points is known to be the optimal precoding strategy. However, finding the optimal precoder has been proved to be NP-hard. For large MIMO, a promising practical approach is to transform the channel into parallel 2 × 2 MIMO subchannels and then precode each of them separately. However, existing methods are mostly based on heuristic subchannel pairing schemes and require numerical search/optimization in the design phase. In this work, we propose a novel real-valued precoder, named as XY-precoder, which enjoys an explicit construction, a provable dmin, a provably optimal subchannel pairing scheme, and low ML-decoding complexity. We prove that the XY-precoder achieves the same diversity order as the best known precoder, but with a much lower decoding complexity. Simulation results confirm that the error performance of XY-precoder is almost the same as that of the best known precoders.

The Unshared Secret Key Cryptography (USK), recently proposed by the authors, guarantees Shannon's ideal secrecy and perfect secrecy for MIMO wiretap channels, without requiring secret key exchange. However, the requirement of infinite constellation inputs limits its applicability to practical systems. In this paper, we propose a practical USK scheme using finite constellation inputs. The new scheme is based on a cooperative jamming technique, and is valid even if the eavesdropper has more antennas than the transmitter. We show that Shannon's ideal secrecy can be achieved with an arbitrarily small outage probability.

The artificial noise (AN) scheme, proposed by Goel and Negi, is being considered as one of the key enabling technology for secure communications over multiple-output multiple-input wiretap channels. However, the decrease in secrecy rate due to the increase in the number of Eve's antennas is not well understood. In this paper, we develop an analytical framework to characterize the secrecy rate of the AN scheme as a function of Eve's SNR, Bob's SNR, the number of antennas in each terminal, and the power allocation scheme. We first derive a closed-form expression for the average secrecy rate. We then derive a closed-form expression for the asymptotic instantaneous secrecy rate with large number of antennas at all terminals. Finally, we derive simple lower and upper bounds on the average/instantaneous secrecy rate that provide a tool for the system design.

While the impact of finite-rate feedback on the capacity of fading channels has been extensively studied in the literature, not much attention has been paid to this problem under secrecy constraint. In this work, we study the ergodic secret capacity of a multiple-input multiple-output multiple-antenna-eavesdropper (MIMOME) wiretap channel with quantized channel state information (CSI) at the transmitter and perfect CSI at the legitimate receiver, under the assumption that only the statistics of eavesdropper CSI is known at the transmitter. We refine the analysis of Lin et al.'s random vector quantization (RVQ) based artificial noise (AN) scheme, where a heuristic upper bound on the secrecy rate loss (compared to the perfect CSI case) was given. We propose a lower bound on the ergodic secrecy capacity. We show that the lower bound and the secrecy capacity with perfect CSI coincide asymptotically as the number of feedback bits and the AN power go to infinity. For practical applications, we propose a very efficient quantization codebook construction method for the two transmit antennas case.

While information-theoretic security is stronger than computational security, it has long been considered impractical. In this work, we provide new insights into the design of practical information-theoretic cryptosystems. Firstly, from a theoretical point of view, we give a brief introduction into the existing information theoretic security criteria, such as the notions of Shannon's perfect/ideal secrecy in cryptography, and the concept of strong secrecy in coding theory. Secondly, from a practical point of view, we propose the concept of ideal secrecy outage and define a outage probability. Finally, we show how such probability can be made arbitrarily small in a practical cryptosystem.

In cryptography, a shared secret key is normally mandatory to encrypt the confidential message. In this work, we propose the unshared secret key (USK) cryptosystem. Inspired by the artificial noise (AN) technique, we align a one-time pad (OTP) secret key within the null space of a multiple-output multiple-input (MIMO) channel between transmitter and legitimate receiver, so that the OTP is not needed by the legitimate receiver to decipher, while it is fully affecting the eavesdropper's ability to decipher the confidential message. We show that the USK cryptosystem guarantees Shannon's ideal secrecy and perfect secrecy, if an infinite lattice input alphabet is used.

Current security techniques can be implemented with either secret key exchange or physical-layer wiretap codes. In this paper, we investigate an alternative solution for MIMO wiretap channels. Inspired by the artificial noise (AN) technique, we propose the unshared secret key (USK) cryptosystem, where the AN is redesigned as a one-time pad secret key aligned within the null space between a transmitter and a legitimate receiver. The proposed USK cryptosystem is a new physical-layer cryptographic scheme, which was obtained by combining traditional network-layer cryptography and physical-layer security. Unlike previously studied AN techniques, rather than ensuring nonzero secrecy capacity, the USK is valid for an infinite lattice input alphabet and guarantees Shannon's ideal secrecy and perfect secrecy without the need for secret key exchange. We then show how ideal secrecy can be obtained for finite lattice constellations with an arbitrarily small outage.

In this paper, a derandomized algorithm for sampling decoding is proposed to achieve near-optimal performance in lattice decoding. By setting a probability threshold to sample candidates, the whole sampling procedure becomes deterministic, which brings considerable performance improvement and complexity reduction over to the randomized sampling. Moreover, the upper bound on the sample size K, which corresponds to near-maximum likelihood (ML) performance, is derived. We also find that the proposed algorithm can be used as an efficient tool to implement soft-output decoding in multiple-input multiple-output (MIMO) systems. An upper bound of the sphere radius R in list sphere decoding (LSD) is derived. Based on it, we demonstrate that the derandomized sampling algorithm is capable of achieving near-maximum a posteriori (MAP) performance. Simulation results show that near-optimum performance can be achieved by a moderate size K in both lattice decoding and soft-output decoding.

Lattice precoding is an effective strategy for multiantenna broadcast. In this paper, we show that approximate lattice precoding in multiantenna broadcast is a variant of the closest vector problem (CVP) known as η-CVP. The proximity factors of lattice reduction-aided precoding are defined, and their bounds are derived, which measure the worst-case loss in power efficiency compared to sphere precoding. Unlike decoding applications, this analysis does not suffer from the boundary effect of a finite constellation, since the underlying lattice in multiantenna broadcast is indeed infinite.

In lattice-coded multiple-input multiple-output (MIMO) systems, optimal decoding amounts to solving the closest vector problem (CVP). Embedding is a powerful technique for the approximate CVP, yet its remarkable performance is not well understood. In this paper, we analyze the embedding technique from a bounded distance decoding (BDD) viewpoint. 1/(2γ)-BDD is referred to as a decoder that finds the closest vector when the noise norm is smaller than λ1/(2γ), where λ1is the minimum distance of the lattice. We prove that the Lenstra, Lenstra and Lovász (LLL) algorithm can achieve 1/(2γ)-BDD for γ ≈ O(2n/4). This substantially improves the existing result γ = O(2n) for embedding decoding. We also prove that BDD of the regularized lattice is optimal in terms of the diversity-multiplexing gain tradeoff (DMT).

Despite its reduced complexity, lattice reduction-aided decoding exhibits a widening gap to maximum-likelihood (ML) performance as the dimension increases. To improve its performance, this paper presents randomized lattice decoding based on Klein's sampling technique, which is a randomized version of Babai's nearest plane algorithm [i.e., successive interference cancelation (SIC)] and samples lattice points from a Gaussian-like distribution over the lattice. To find the closest lattice point, Klein's algorithm is used to sample some lattice points and the closest among those samples is chosen. Lattice reduction increases the probability of finding the closest lattice point, and only needs to be run once during preprocessing. Further, the sampling can operate very efficiently in parallel. The technical contribution of this paper is twofold: we analyze and optimize the decoding radius of sampling decoding resulting in better error performance than Klein's original algorithm, and propose a very efficient implementation of random rounding. Of particular interest is that a fixed gain in the decoding radius compared to Babai's decoding can be achieved at polynomial complexity. The proposed decoder is useful for moderate dimensions where sphere decoding becomes computationally intensive, while lattice reduction-aided decoding starts to suffer considerable loss. Simulation results demonstrate near-ML performance is achieved by a moderate number of samples, even if the dimension is as high as 32.

Sphere decoding achieves maximum-likelihood (ML) performance at the cost of exponential complexity; lattice reduction-aided decoding significantly reduces the decoding complexity, but exhibits a widening gap to ML performance as the dimension increases. To bridge the gap between them, this paper presents randomized lattice decoding based on Klein's randomized algorithm, which is a randomized version of Babai's nearest plane algorithm. The technical contribution of this paper is two-fold: we analyze and optimize the performance of randomized lattice decoding resulting in reduced decoding complexity, and propose a very efficient implementation of random rounding. Simulation results demonstrate near-ML performance achieved by a moderate number of calls, when the dimension is not too large.