In indoor ultra-wideband (UWB) positioning systems, accuracy can be improved by determining the conditions of line-of-sight (LOS) and non-line-of-sight (NLOS) propagation taking appropriate measures. The existing methods, such as support vector machine (SVM), decision tree (DT), k-Nearest Neighbor (KNN), identify LOS/NLOS mainly using time-domain characteristics. However, only characteristics cannot achieve satisfactory performance. this letter, we propose a identification method based on...

Embedding-based retrieval models have made significant strides in retrieval-augmented generation (RAG) techniques for text and multimodal large language (LLMs) applications. However, when it comes to speech larage (SLLMs), these methods are limited a two-stage process, where automatic recognition (ASR) is combined with text-based retrieval. This sequential architecture suffers from high latency error propagation. To address limitations, we propose unified embedding framework that eliminates...

The accuracy of positioning in ultra-wide band (UWB) systems can be improved by identifying the conditions line-of-sight (LOS) and non-line-of-sight (NLOS) propagation taking appropriate measures. At present, machine learning methods have already been extensively applied to NLOS identification. Unfortunately, those traditional are incapable performing well cases small samples class-imbalance samples. In this letter, a novel method is proposed identify LOS components utilizing Capsule...

This paper studies a multi-unmanned aerial vehicle (UAV) enabled wireless communication system, where multiple UAVs are employed to communicate with group of ground terminals (GTs) in the presence potential jammers. We aim maximize throughput overall GTs by jointly optimizing UAVs' trajectory, GTs' scheduling and power allocation. Unlike most prior studies, we consider turning climbing angle constraints, three-dimensional (3D) trajectory minimum UAV-to-UAV (U2U) distance constraint, transmit...

This letter studies an unmanned aerial vehicle (UAV)-enabled uplink communication system with turning and climbing angle constraints, the process suffers from attack one ground jammer imperfect location information. Under such setup, we aim to maximize throughput by jointly considering UAV's three dimensions (3D) trajectory design node (GN)'s transmit power control. However, optimization problem is non-convex hard be tackled. To overcome this difficulty, develop alternating algorithm...

The integration of sensing and communication (ISAC) model is regarded as a key enabling technology for B5G 6G, has gained growing attention from both academia industry. This letter proposes Multi-domain NOMA (MD-NOMA)-based ISAC scheme. proposed scheme enables parallelly non-orthogonally transfer data streams in the TF (Time-Frequency) DD (Delay-Doppler) domains while potential reflectors. additional DoF provided by domain utilized effectively. An iteration-based receiver to mitigate...

High-quality semen with high viability is critical to improving the in-vitro fertilization efficiency. This study aimed understand effect of ambient temperature and humidity on quality seminal plasma biochemical parameters Mediterranean buffalo in March July. The metabolites two seasons were detected using UPLC-MS/MS method. results showed that index (THI) 66.86 ± 2.98, 82.94 3.52 Compared March, breath frequency, rectal temperature, heat shock protein 70 expressions significantly increased...

The thin film narrow band-pass filters are widely used in precision optical systems such as gas detection, fiber optic communication and analysis instruments due to their unique characteristics. However, the applications become more widespread, problems of regulating half bandwidth cut-off based on Fabry-Perot (F-P) principle becoming apparent. In this work, we took filter with a center wavelength 450 nm design target, comprehensively optimized system standard F-P obtain different bandwidths...

Cross-Camera vehicle tracking(CCVT) is an significant field in intellective transportation systems and intelligent visual perception smart city. It challenging due to the large variation of same within various cameras. We propose a cross-camera multi-target tracking framework towards traffic flow parameter estimation, which mainly divided into three steps:(i) multi-object single camera, (ii) multi-camera trajectory association and, (iii) estimation. Among them, we design encoder that...

X-ray baggage security screening is a demanding task for aviation and rail transit security; automatic prohibited item detection in images can help reduce the work of inspectors. However, as many items are placed too close to each other baggages, it difficult fully trust results intelligent algorithms. In this paper, human-in-the-loop inspection framework proposed. The proposed utilizes deep-learning-based algorithm find suspicious images, select manual examination when cannot determine...

This study proposes a vision‐based method for traffic sign attribute estimation, i.e. 3D position and pose, from image sequences by binocular or monocular cameras. The starts with acquiring robust feature correspondences based on homography constraints pairs. Then the objective function is designed to integrate optimise parameters of plane in coordinate. Finally, utilised estimation. In addition, authors provide an extension raw KITTI dataset, which can be tasks localisation pose...

X-ray security automatic screening has attracted a broad attention for providing auxiliary support to eliminate potential threat in public. However, recent detection algorithms may not perform quite well when great number of dangerous articles are overlapping and compact the images. In order deal with it, this paper takes segmentation scheme network into consideration. Based on Mask R-CNN, we propose two-stage deep CNN model called SAN solve performance degradation caused by dense...

This paper studies the anti-jamming issue for cognitive unmanned aerial vehicle (UAV) communication networks. We aim to maximize throughput of information collection in sensitive areas or military borders by optimizing UAV trajectory and secondary transmitters (STs)' transmit power, presence malicious jammers with imperfect location information. However, optimization problem is challenging solve due non-convexity. To overcome this difficulty, we propose a maximization algorithm aid block...

Road scene model construction is an important aspect of intelligent transportation system research. This paper proposes framework that can automatically construct road models from image sequences. The and foreground regions are detected at superpixel level via a new kind random walk algorithm. seeds for different initialized by trapezoids propagated adjacent frames using optical flow information. region detection implemented the algorithm, which then refined fast two-cycle set method. After...

In the tasks of multi-robot collaborative area search, we propose unified approach for simultaneous mapping sensing more targets (exploration) while searching and locating (coverage). Specifically, implement a hierarchical multi-agent reinforcement learning algorithm to decouple task planning from execution. The role concept is integrated into upper-level selection, which enables robots learn based on state status upper-view. Besides, an intelligent switching mechanism selection module...

Learning robust driving policy for complex scenarios brings significant challenges to autonomous system due its inherent black-box properties. Recent methods learn policies from not only ego-vehicle but all the observed vehicles improve sample efficiency and scenario exposure. However, our empirical studies unveil existence of meaningless erroneous behaviors vehicles, which may result in unexpected dangerous policies. Expanding upon notion influence function, gauges effect a training on...

Since public transportation scenarios have high requirements for baggage security, accurate X-Ray Object Segmentation security inspection is vital to help maintain social security. images generated by different X-ray machines differ in terms of color schemes, pixels and geometric deformation, detection segmentation methods trained on certain datasets generalize poorly out-of-distribution images. Recently, the Segment Anything Model (SAM) has been published as a foundation model image...

In this paper, a new multi-modal traffic scene simulation framework with combined inputs of road image sequences and information from Geographic Information Systems (GIS) is proposed. The proposed contains two major steps, the first one being preprocessing step, including 3D model extraction, camera location orientation estimation lane extraction both GIS sequences. After such preprocessing, reconstruction reformulated into 6-degree freedom (6DoF) pose in model. Subsequently, iterative...

This paper investigates the influence of uncertainty jammers' positions and UAV's turning angle constraint jointly by studying a UAV-enabled wireless communication system, where UAV is dispatched to collect information from group ground terminals (GTs) in presence potential jammers. We aim maximize average throughput all GTs optimizing trajectory GTs' scheduling. The formulated problem shown be mixed integer non-convex optimization that intractable work out. Thus, we develop robust...

This study addresses the problem of non‐perspective pose estimation from line correspondences in traffic scenarios. A coarse‐to‐fine 3D road registration method is proposed for this two stages. Firstly, iterative closest point algorithm exploited to estimate coarsely. An objective function then established incorporate feature refining coarse pose. Besides, framework including employed video augmentation. The begins with inputs videos, information Geographic Information Systems and models...