Smart speakers allow users to interact with home appliances using voice commands and are becoming increasingly popular. While voice-based interface is intuitive, it insufficient in many scenarios, such as noisy or quiet environments, for language barriers, applications that require continuous motion tracking. Motion-based control attractive complementary existing control. However, accurate reliable room-scale tracking poses a significant challenge due low SNR, interference, varying mobility....

In this article, a new hybrid deep learning (DL) algorithm is developed to make computer-assisted forecasting energy management (EM) system. Applying the Copula function, Hankel matrix created for processing gathered automatic metering infrastructure (AMI) load information in smart network. This of data results model optimization through suggested pooling-based neural network (PDNN). Through increased size and variation AMI data, PDNN reduces overfitting issues during testing training. The...

Indoor localization based on RSS fingerprinting approach has been attracting many research efforts in the past decades. Recent study presents a fundamental limit of approach: given requirement estimation accuracy, reliability user's result can be derived. As highly accurate indoor is essential to enable location services, natural question ask is: we further improve accuracy scheme fundamentally? In this paper, theoretically show that temporal correlation localization. particular, construct...

We present a spatiotemporal model, namely, procedural neural networks for stock price prediction. Compared with some successful traditional models on simulating market, such as BNN (backpropagation networks, HMM (hidden Markov model) and SVM (support vector machine)), the network model processes both spacial temporal information synchronously without slide time window, which is typically used in well‐known recurrent networks. Two different structures of are constructed modeling...

For existing indoor localization algorithm has low accuracy, high cost in deployment and maintenance, lack of robustness, sensor utilization, this paper proposes a particle filter based on multi-sensor fusion. The pedestrian’s environment is described as dynamic system state estimation problem. combines the smart mobile terminal with localization, filters result filter. In paper, interval pedestrian dead reckoning (PDR) information RSSI have been used to improve filtering precision...

Aiming at the shortcomings of existing indoor location algorithm, such as low accuracy positioning, high deployment and maintenance cost, unstable robustness, this paper proposes a method based on integration smartphone with WiFi magnetic field using multi-sensor fusion. In initial stages rough is achieved by Wi-Fi-RSSI fingerprints which provides an geomagnetic matching area for positioning particle filter matching. This use median algorithm to deal original data covariance interpolation...

Corrosion defects will increase the risk of power equipment failure, which directly affect stable operation systems. Although existing methods can detect corrosion equipment, these are often poor in real-time. This study presents a two-stage detection approach that combines YOLOv8 and DDRNet to achieve real-time precise area localization. In first stage, network is used identify locate substation detected ROI areas passed second stage for semantic segmentation. To enhance performance both...

The magnetic information measured on the smartphone platform has a large fluctuation and research of indoor localization algorithm based smart-phone is less. Indoor particle filter studied. Robust local weighted regression used to smooth original data in process constructing map. Use moving average filtering model online observation positioning process. Compare processed with map collected by matching error 0.3941uT. Average 0.229 meter when using data.

Efficient localization algorithm is a research focus in the field of wireless sensor network (WSN). Nowadays, for based on support vector regression(SVR) one most important algorithms. In this paper, we analyse model regression and propose novel extraction method training data model. By simulations, results show proposed gives better accuracy than original programming localization.

As a core procedure in an increasingly automated industrial environment, defect detection is important for producing micro armatures. However, there are still some difficulties using general deep learning methods, the spatial feature loss problem small targets of solder surface defects and difficulty tuning reference sample selection strategy. Therefore, this paper proposes adaptive label assignment framework. The detector uses anchor free object network as backbone. A compensation method...

Indoor human action recognition, essential across various applications, faces significant challenges such as orientation constraints and identification limitations, particularly in systems reliant on non-contact devices. Self-occlusions non-line of sight (NLOS) situations are important representatives among them. To address these challenges, this paper presents a novel system utilizing dual Kinect V2, enhanced by an advanced Transmission Control Protocol (TCP) sophisticated ensemble learning...

Active electronically scanned array antenna (AESA antenna) is capable of controlling the radiation pattern by feeding phase radiating elements. It has good performance and plays an important role in radar systems. With development AESA towards high-frequency bands high-density arrays, structural-electromagnetic-thermal (SET) coupling becomes increasingly significant. seriously restricts realization high performances antennas. However, previously reported theoretical multi-field-coupled model...

The ultrawideband (UWB) indoor positioning system has poor accuracy and stability in the environment with complex structure, which is affected by non-line-of-sigh (NLOS), multi-path effect, physical measurement other factors. Therefore, this paper proposes a UWB base station (BS) selection method based on time difference of arrival (TDOA) fingerprint. Firstly, TDOA information BS reference point are collected off-line phase. Secondly, on-line phase, improved reversal resonating strength...

In order to improve accuracy and reliability of wireless location in the mixed line-of-sight (LOS) non-line-of-sight (NLOS) environment, a real-time NLOS identification LOS reconstruction algorithm connection with positive characteristic error are presented estimate position mobile station (MS). A Kalman filter (KF) based on innovation vectors is utilized identify change measured value between LOS, which identification. According deviation can be as part error, original estimation curve by...

For indoor sensor systems, it is essential to implement an extra supporting area notification part. To inform the real-time coordinates, time difference of arrival (TDOA) algorithm can be introduced. these localization their main processes are often built based on line sight (LOS) scenario. However, obstacles make off-the-shelf system unable play its due role in flexible non-line (NLOS) So, necessary adjust signals according NLOS identification results. methods before were not effective...

The demand for indoor positioning is exploding. For positioning, TPSN ranging model with specific frequency band often used based on android platform or others. Time of arrival (TOA) as an important part to achieve efficient localization. signal wav files classification line sight (LOS) and non-line (NLOS) will be involved. Support Vector Machine (SVM) complete in most systems the effect not so good. At this moment, order improve performance system, Shrinkage Enhanced Particle Swarm...