To address the limitations of conventional radar systems in complex terrains, this paper focuses on two critical challenges. The first challenge is that traditional Synthetic Aperture Radar (SAR) inherently struggles with persistent monitoring due to discontinuous observation patterns and rigid imaging geometries. The second challenge is that fixed phased-array radars lack dynamic adaptability to irregular terrain surfaces despite their beam steering capability. To overcome these constraints, we propose a distributed phased-array radar system using collaborative Unmanned Aerial Vehicle (UAV) swarms. By deploying Frequency-Modulated Continuous Wave Multiple-Input Multiple-Output (FMCW-MIMO) radar modules on autonomous UAVs, the system constructs a reconfigurable virtual array that enables adaptive 3D aperture synthesis. The core innovation lies in the synergistic integration of motion compensation algorithms and distributed synchronization mechanisms, which resolves the inherent conflict between platform mobility and observation stability in conventional radar architectures. Experimental validation confirms enhanced performance in complex terrains, achieving superior angular resolution and coverage efficiency compared to fixed-phased arrays.
Large language model (LLM) applications deployed in regulated and high-risk environments face two coupled runtime risks: input-side prompt injection and output-side content compliance violations. Existing defenses are often either model-centric (alignment or fine-tuning) or application-local (keywords and regex), which limits auditability, policy agility, and cross-model portability. This paper proposes a gateway-centric framework that externalizes safety enforcement into a security side-chain. The framework integrates a unified LLM gateway with an external semantic guardrail service to perform bidirectional inspection before and after model invocation. We formalize the architecture with input-risk and output-compliance functions, define a comprehensive threat model and trust boundary, and present a reproducible evaluation protocol. To validate our approach, a prototype implementation based on LiteLLM and Prisma AIRS was evaluated on a rigorously balanced 2,000-sample adversarial dataset merged from diverse sources, including Deepset, wildjailbreak, and jailbreak-classification. Results demonstrate that the proposed architecture achieves an exceptional balance in detection capability, yielding an F1-score of 0.8322, a recall of 0.8531, and effectively maintaining a low False Positive Rate (FPR), thereby outperforming multiple state-of-the-art baselines including Llama Prompt Guard and Deepset DeBERTa. Furthermore, performance testing confirms practical deployability, showing an average added latency of only 320 ms under a 50-concurrent-user load. The paper concludes by discussing the side-chain advantages over model alignment mechanisms, guardrail bypass risks, and adaptive policy-evolution strategies for enterprise AI governance.
In the field of forensic document examination, the non-destructive and precise differentiation of signing pen inks is pivotal for determining document authenticity and reconstructing case facts. Hyperspectral imaging (HSI), characterized by the integration of spatial and spectral information, has become a vital tool for ink analysis; however, the massive volume of high-dimensional data and significant inter-band redundancy pose challenges for traditional classification algorithms. Herein, we propose integrating a convolutional neural network (CNN) and a Kolmogorov-Arnold network (KAN) into a hybrid CNN-KAN architecture that employs the CNN as its front-end to extract local hierarchical features and compress dimensions and the KAN module as its back-end to enhance non-linear fitting capabilities, thereby effectively discriminating between highly similar ink categories. This approach addresses the limited fitting capacity of standalone CNNs and mitigates the excessive computational costs associated with applying pure KANs to high-dimensional data. Experimental results demonstrate that in a classification task involving 40 brands and models of signing pen inks, the CNN-KAN model achieved an accuracy of 98.56%, outperforming traditional CNN, Visual Geometry Group (VGG), feedforward neural network (FNN), and attention-guided U-shaped network (AUNet) models. Furthermore, compared to a pure KAN, the proposed hybrid architecture reduced the computational load by approximately 60%, achieving a balance between high precision and computational efficiency. This study provides a superior deep learning solution for forensic evidence identification based on HSI technology.
Visual speech recognition (VSR) recovers speech from lip movements when audio is noisy or unavailable. Its multi-stage inference pipeline spans video decoding, mouth-region extraction, preprocessing, model invocation, and decoding, where upstream failures can mask downstream faults. Pipeline maintenance therefore still relies largely on predefined checks and manual debugging. We propose LoopVSR, a Loop Engineering framework that enables a code agent to automatically diagnose and repair VSR inference pipelines using end-to-end execution evidence. It couples constrained repository-level diagnosis and patching with an external controller that audits changes, runs real inference, and accepts or rolls back patches using failures and character error rate (CER). The resulting feedback loop returns newly observed exceptions, tensor statistics, and recognition errors to the agent, progressively exposing faults masked by upstream failures. On the CMLR VSR system, LoopVSR repairs all 11 main faults with 100
This paper presents a remote automatic water quality monitoring system based on the STM32 microcontroller and wireless sensor network (WSN) technologies. The system integrates a hybrid communication architecture combining ZigBee for short-range multi-hop networking and General Packet Radio Service (GPRS) for long-range data transmission via a Machine-to-Machine (M2M) platform, enabling real-time synchronous monitoring of four key water quality parameters: pH, fluoride, electrical conductivity (EC), and temperature. A solar dual-panel power supply module ensures long-term autonomous field operation without external power dependence. Additionally, an automatic water sample retention mechanism preserves samples when any parameter exceeds preset thresholds for subsequent laboratory analysis. The results show that the average relative detection errors for pH, fluoride, EC, and temperature are −2.8%, −1.41%, −3.02% and −1.89%, respectively, all within $\pm \mathbf{5 {\%}}$. This system features low cost, high precision, solar-powered autonomy, sample retention beyond limits, and immediate warning capability, offering significant practical value for extensive deployment in water quality monitoring applications.
