In multi-target tracking, a single model is insufficient to ensure the stable tracking of the maneuvering targets. This paper proposes a multi-model Probability Hypothesis Density (PHD) filtering to address the coexistence of point and extended target (MM-PE-PHD). The multi-model PHD filter utilizes Jump Markov System (JMS) to extend a single motion model to multiple motion models. In most realistic tracking scenarios, the motion model of the target is usually time-varying, and there is often a coexistence of the point target and the extended target. Therefore, it is necessary to perform an adaptive estimation of the target model and build a generalized measurement model to track the coexistence point and extend target. In the MM-PE-PHD algorithm of this paper, the model probabilities of each target and the probability that the target is a point target are also estimated concurrently. To obtain a closed-form solution, the intensity function of PHD is modeled as a mixture of the Gaussian distribution and the Gamma Gaussian Inverse Wishart (GGIW) distribution. Combined with the probability of the point target and model probability, the type of the target, the motion model, the kinematic state, and the extended state are all estimated simultaneously. The simulation experiments validated the effectiveness and superiority of the proposed algorithm.

A lot of progress has been made in sampling theory, with numerous review papers and books having been published on it and its applications. In these situations, 1-D nonuniform samplings, where several different sampling time periods in rational ratios are used, have been intensively discussed. In this paper, Proposition 1 shows that such samplings are found equivalent to kinds of higher-order samplings where sampling positions are expressed using one common sampling time period with several different sampling offsets. Next, 1-D nonuniform samplings, where several different sampling time periods in irrational ratios are used, are discussed. Proposition 2 reveals that such nonuniform samplings do not exist, as positioning of the sampling points is not feasible, by using Kronecker’s theorem. The results obtained from Propositions 1 and 2 indicate that all the feasible nonuniform samplings are equivalent to kinds of higher-order samplings. Then, the necessary and sufficient condition for a given 1-D signal to be sampled satisfying fidelity and optimality in higher-order samplings is examined. Proposition 3 indicates that the answer is that its frequency support is made up of tiling clusters of one common lattice in frequency space, which is reciprocally related to a certain lattice in time space. This result can be, of course, applied to the nonuniform samplings equivalent to them.

A term tree pattern is a rooted ordered tree pattern which consists of ordered tree structures with edge labels and structured variables with labels. A variable in a term tree pattern is a labeled hyperedge which can be replaced with any ordered tree that has at least 2 vertices. In this paper, we focus on a special type of variable, called a contractible variable. A contractible variable can be replaced with any ordered tree that has at least 1 vertex. A usual variable, called an uncontractible variable, cannot be replaced with an ordered tree that consists of a single vertex. The matching problem for term tree patterns is to decide whether or not a given ordered tree can be obtained from a given term tree pattern by replacing variables in the term tree pattern with ordered trees. Previously, we proposed a polynomial time matching algorithm for solving the matching problem for term tree patterns having uncontractible variables. In this paper, we propose a matching algorithm for solving the matching problem for term tree patterns having contractible variables and uncontractible variables. Furthermore, we show that the proposed algorithm correctly solves this problem in O(n⁴ N) time, where n and N denote the numbers of vertices of a given term tree pattern and a given ordered tree, respectively.

This paper proposes an error rate analysis method for Binary Ring Learning with Errors (bin-RLWE) cryptosystems where the secret and error are binary. We derive a computable method for the bit error rate (BER) in encryption and decryption processes using a normal approximation, and estimate an upper bound on the overall decryption failure probability (at least one bit error) via a union bound. The framework was validated through 10⁹ full encrypt-decrypt tests for each of four parameter sets. Based on this analysis, we provide parameter selections for implementations requiring overall failure probability bounds of 2^-16 and 2^-32. We also evaluate the hardware performance of these parameters on 28 nm ASIC and Xilinx Kintex-7 FPGA platforms, measuring resource usage and latency. Comparisons with conventional RLWE implementations show that bin-RLWE achieves better resource efficiency and faster processing. Unlike Number Theoretic Transform-based RLWE implementations that commonly constrain dimensions in practice, our shifting convolution-based approach removes this implementation constraint, enabling more flexible parameter choices. Finally, we discuss the correctness-security trade-off in bin-RLWE and report estimated security levels against primal attack for the selected parameter sets.

Messenger, a messaging service, is widely used in the United States and around the world, with 947 million monthly active users as of February 2025. In December 2023, a default implementation of end-to-end encryption in Messenger and two accompanying technical white papers were released. Labyrinth Encrypted Message Storage Protocol is one of the protocols presented in these technical white papers and is intended to securely share message history between devices associated with each Messenger account. However, there has been no rigorous discussion of the security of Labyrinth to date, and the security model claimed in the technical white paper has some ambiguity. In this paper, we give the first formal verification of Labyrinth. First, we verify the confidentiality and integrity of the stored messages as the security claim in the specification (i.e., only considering outside adversaries). Next, considering the case of a device that has been revoked from an account and abused, we exhaustively verify under what conditions it is possible to break the confidentiality and integrity of a message. As a result of the verification, we clarify the conditions under which the confidentiality and integrity of stored messages can be broken by abusing a revoked device and the specific procedures for doing so while satisfying all the security claims made by the specification.

Distributed matrix multiplication (DMM) has broad applications in data analytics, machine learning, and scientific simulation, but faces critical challenges such as data privacy concerns, server straggling, and Byzantine node failures. To address these issues, this paper proposes three novel secure distributed matrix multiplication (SDMM) schemes. The first scheme utilizes Reed-Solomon (RS) codes and their duals to achieve resilience against stragglers, security against Byzantine servers and information-theoretic privacy against colluding servers. By leveraging the orthogonality of roots of unity, it supports flexible grid partitioning for matrices A and B. To the best of our knowledge, this is the first root-of-unity-based SDMM construction that offers general straggler resilience under grid partitioning and achieves a significantly lower recovery threshold than prior schemes. To further enhance communication efficiency, the second scheme employs subspace polynomials to reduce the per-server download cost by a factor of s/ℓ while preserving the same privacy guarantees; subsequently, a lower bound on the download cost for communication efficiency (CE)-SDMM schemes is derived. Finally, inspired by RS codes over cyclic polynomial rings, the third scheme replaces costly finite-field multiplications with simple cyclic shifts and XOR operations over the rings, thereby reducing the computational complexity. Moreover, it can be efficiently implemented using the developed fast Fourier transform (FFT) algorithms, further reducing the overall encoding and decoding complexity.

With the rapid development of location-based services, indoor positioning has become a research hotspot. Channel State Information (CSI) has emerged as a promising signal source for high-precision indoor localization. However, existing methods still face challenges in robustness and accuracy. This paper proposes a WiFi CSI fingerprinting model that performs outlier removal and denoising in the offline stage. After processing, the amplitude and phase of CSI are fused and input into a CNN-Transformer network to capture spatial features and global dependencies. To address the difficulty of hyperparameter tuning in Transformers, an Improved White Shark Optimization (IWSO) algorithm is introduced to optimize key parameters and enhance performance. In the online stage, positioning is achieved through real-time CSI measurement. Experimental results show that the proposed method improves positioning accuracy by 18.33% compared to models without hyperparameter tuning, achieving an average positioning error of 1.35 m.

In this study, a novel model called CFE-UNet has been proposed for extracting the context feature to improve the model performance with moderate parameters and computational cost in the field of medical image segmentation. The proposed model comprises two core components, which are context-aware module and hierarchical neighborhood feature extraction module, respectively. These components are instrumental in context feature extraction, thereby enhancing the model’s performance. In addition, a dataset comprising DWI images of acute ischemic stroke (AIS) has been proposed for testing in this study. Extensive experiments have been conducted on the proposed dataset together with four other commonly used datasets, which are ISIC 2018, BUSI, GlaS, and Kvasir-SEG, respectively. The promising results demonstrate that the proposed CFE-UNet model achieves exceptional performance with moderate model parameters and computational costs.

The least mean square algorithm has attracted a lot of attention among the various adaptive algorithms due to its simple structure and good computational complexity. The step-size is a key factor affecting some critical performance and the variable step-size strategy has become an important branch of research. However, some methods lack systematic theoretical derivation and appropriate analysis. The difficulty of obtaining prior parameters and the common performance make some algorithms less practical. Therefore, a variable step-size algorithm which is based on the energy of the posterior error without noise and combined with a dynamic restricted function is proposed in this paper. This algorithm takes the energy of the noiseless posterior error as the cost function and uses the basic theoretical formula to obtain the variable step-size expression. The parameters of the proposed algorithm can be obtained by an iterative smooth calculation without requiring any prior information. The specific mean square deviation performance analysis of the algorithm is also provided. Various experimental results show that the proposed algorithm can accurately adjust the step-size according to the convergence state of the system.

With the rapid advancement of smart cities, the demand for accurate indoor positioning is increasing. Traditional methods face challenges from high data collection costs and multipath effects. This letter proposes an Enhanced Attention Conditional GANs (EA-CGAN) based on fingerprinting for few-sample scenarios. By integrating SE and CBAM attention mechanisms, it optimizes network architecture and enhances spatial and channel feature extraction. EA-CGAN generates high-quality augmented RSS fingerprint databases from CGAN class labels, even with limited data. Experiments show that with only 20% real data, it achieves localization accuracy comparable to conventional methods, significantly reducing costs.

Zhang et al. (IEEE Internet of Things Journal (SJR Q1 journal) 2025) proposed multi-user keyword-field-free conjunctive searchable encryption (MU-KFFCSE). They claimed that the MU-KFFCSE scheme provides Trapdoor Indistinguishability (TI) where no keyword information is revealed from a trapdoor. In this short note, we demonstrate that the MU-KFFCSE scheme does not provide TI security.

A vanity address is a blockchain address that incorporates custom characters, such as a name. Traditionally, vanity addresses are generated using a classical try-and-error (i.e., a brute-force) approach, which limits the number of characters that can be embedded. In this short note, we explore the potential of identity-based signatures (IBS), which allow arbitrary strings to be used as verification keys, to facilitate the generation of vanity addresses. We focus on a generic construction of IBS from conventional signature schemes and propose an IBS scheme based on ECDSA with key recovery. Although direct generation of vanity addresses is not feasible due to the key recovery mechanism inherent in ECDSA, the IBS framework enables any string to be associated with an address, thereby imparting additional semantic meaning to it. We implement our system in Solidity and demonstrate that the gas cost is nearly equivalent to that of standard ECDSA signature verification.

This letter presents a new class of perfect Gaussian integer sequences (PGISs), called Constant Euclidean-norm PGISs, where every Gaussian integer has the same Euclidean norm. Several proposed construction methods are based on the combination of different base sequences. With the aid of Pythagorean theorem, distinct elements appearing in the base sequences result in many new families of constant Euclidean-norm PGISs with degree of freedom. Both theoretical guarantees and numerical examples of the newly proposed construction methods are provided as well.

Binary sequences with low correlation properties have wide applications in communications systems. Recently, binary sequences of period 4n with optimal autocorrelation are proposed by interleaving four suitable Ding-Helleseth-Lam sequences (Des. Codes Cryptogr., https://doi.org/10.1007/s10623-017-0398-5), where n is an odd prime. The objective of this letter is to discuss the cross-correlation property of optimal binary sequences of length 4n by using interleaving technique. It turns out that eight binary sequence pairs have low cross-correlation magnitude upper bounded by 4$\sqrt{n-4}$. Those proposed sequence pair could be used for synchronization in communication systems.

To reduce the multi-access interference (MAI) among different users near the signal origin in quasi-synchronous frequency hopping multiple access (QS-FHMA) communication systems, low-hit-zone frequency hopping sequence (LHZ-FHS) has been widely applied. The WG-FHS can improve the resistance to narrowband interference, broadband interference, follower jamming, frequency blocking interference, multipath fading, etc. In this paper, a new class of WG-LHZ-FHS set is presented using the interleaving technique. It satisfies the wide-gap characteristic and is optimal with respect to the Peng-Fan-Lee bound.

PCMNoise uses standardized phase-change memory conductance readouts as weight noise for exploration in deep RL. The hardware noise is distributionally equivalent to NoisyNet and validated across 10 conductance states. On Atari and MuJoCo, PCMNoise improves final performance and cuts randomness-related training energy by 15.6× versus PRNGs.