Cluster analysis constitutes a pivotal component of database mining, with DBSCAN being one the most extensively employed algorithms in this domain. Nevertheless, is encumbered by several limitations, including challenges processing high-dimensional datasets, pronounced sensitivity to input parameters, and inconsistencies generating reliable clustering outcomes. This paper presents refined version that utilizes block-diagonal property similarity graphs enhance process. The core concept...

Radio map construction requires a large amount of radio measurement data with location labels, which imposes high deployment cost. This paper develops region-based from received signal strength (RSS) measurements without labels. The is based on set blindly collected RSS device that visits each region in an indoor area exactly once, where the footprints and timestamps are not recorded. main challenge to cluster match clusters physical regions. Classical clustering algorithms fail work as...

Subspace-based models have been extensively employed in unsupervised segmentation and completion of human motion sequence (HMS). However, existing approaches often neglect the incorporation temporal priors embedded HMS, resulting suboptimal results. This paper presents a subspace variety model for along with an innovative Temporal Learning Subspace Variety Model (TL-SVM) method enhanced HMS. The key idea is to segment incomplete HMS into clusters extracting features each through learning...

Water pollution has been hindering the world's sustainable development. The accurate inversion of water quality parameters in sewage with visible-near infrared spectroscopy can improve effectiveness and rational utilization management resources. However, accuracy spectral models is usually prone to noise information high dimensionality data. This study aimed enhance model through optimizing based on sensitive intervals different parameters. To this end, six kinds taken from a biological...

This paper proposes an RSS-based approach to reconstruct vehicle trajectories within a road network, enforcing signal propagation rules and mobility constraints mitigate the impact of RSS noise sparsity. The key challenge lies in leveraging latent spatiotemporal correlations data while navigating complex networks. To address this, we develop Hidden Markov Model (HMM)-based embedding (HRE) technique that employs alternating optimization infer from measurements. model captures dependencies...

Elevators are essential tools in daily life; timely and accurate fault diagnosis plays a crucial role ensuring their safe operation. However, the existing elevator methods often neglect imbalance between actual collected normal samples samples, resulting low diagnostic accuracy. In this study, we propose an improved Aquila optimizer (IAO) extreme gradient boosting tree (XGBoost)-based method under unbalanced samples. The proposed includes three main components: multi-domain feature...

Indoor localization is important for many location- based services. The fundamental challenge the high deployment cost device, infrastructure, and calibration. This paper develops a blind calibration approach received signal strength (RSS)-based localization. essential idea to employ device that visits each region exactly once in an indoor area complete data collection process without recording route, locations, timestamps. Thus, key cluster training into groups extract features identify...

This paper studies the problem of predicting spectrum efficiency (SE) for massive multiple-input multiple-output (MIMO) empowered 5G networks based on reference signal received power (RSRP) collected from drive test (DT). is challenging because there no precise model between RSRP and SE. The SE not only depends RSRP, which captures statistic channel, but also beamforming strategy serving base station (BS) interference neighboring cells, are measured at client. adopts a model-assisted...

The three-dimensional (3D) models of buildings and plants from UAV images become increasingly popular for city construction. However, whether the previous 3D modeling precision large-scale can be further enhanced that is acceptable still remain to investigated. For these ends, this research studied a basketball hall row Euonymus japonicas based on DJI Inspire-1 system. data were processed with Pix4D calculate camera parameters, which then ContextCapture Photoscan generate models. displayed...

Constructing channel state information (CSI) maps may help wireless communications and localization. However, CSI map construction requires up-to-date measurement data with location labels, which induces a huge challenge in practice. Conventional embedding methods project the to low dimensional latent space not have clear physical meaning for localization purpose. This paper attempts extract user locations from measurements recover trajectory of an outdoor vehicular communication scenario. A...

Radio map construction requires a large amount of radio measurement data with location labels, which imposes high deployment cost. This paper develops region-based from received signal strength (RSS) measurements without labels. The is based on set blindly collected RSS device that visits each region in an indoor area exactly once, where the footprints and timestamps are not recorded. main challenge to cluster match clusters physical regions. Classical clustering algorithms fail work as...

In order to carry out effective anti-reconnaissance for unmanned aerial vehicles (UAV), the feasibility of monitoring UAV by laser detection system has been analyzed. The models transmitting and receiving photoelectric imaging on are established. characteristics specular scattering focal plane reflection lens simulated. Taking located at a height 10,000 m as an example, application two systems under reflected light range incident angle laser, distribution intensity photosensitive surface...