- Bayesian Modeling and Causal Inference
- Target Tracking and Data Fusion in Sensor Networks
- Traffic Prediction and Management Techniques
- Sports Analytics and Performance
- Data Quality and Management
- Statistical Methods and Bayesian Inference
- Data Management and Algorithms
- Underwater Vehicles and Communication Systems
- Fault Detection and Control Systems
- Radar Systems and Signal Processing
- Indoor and Outdoor Localization Technologies
- Distributed Sensor Networks and Detection Algorithms
- Stock Market Forecasting Methods
- Anomaly Detection Techniques and Applications
- Traffic control and management
- Time Series Analysis and Forecasting
- Maritime Navigation and Safety
- Transportation Planning and Optimization
- Advanced Neural Network Applications
- Advanced Measurement and Detection Methods
- Video Surveillance and Tracking Methods
- Advanced Decision-Making Techniques
- Bayesian Methods and Mixture Models
- Plant nutrient uptake and metabolism
- Data Mining Algorithms and Applications
Shanghai Maritime University
2011-2024
Henan Agricultural University
2023-2024
Verizon (United States)
2024
Institute of Geographic Sciences and Natural Resources Research
2023
Chinese Academy of Sciences
2023
Xidian University
2020-2023
Nanjing University of Posts and Telecommunications
2014-2022
Peking University
2022
Institute of Electronics
2021
Dalian Neusoft University of Information
2020
Among the current indoor positioning technologies, Bluetooth low energy (BLE) has gained increasing attention. In particular, traditional distance estimation derived from aggregate RSS and signal-attenuation models is generally unstable because of complicated interference in environments. To improve adaptability robustness BLE system, we propose making full use three separate channels instead their combination, which been used before. first step, are separately established for each...
Abstract This article explores recursive algorithms for parameter identification issues of Hammerstein output‐error systems. The proposed approach includes the key term separation auxiliary model gradient algorithm, which utilizes search and separation. To enhance computational efficiency, system is decomposed into two or three subsystems through hierarchical principle. Based on this, a based two‐stage algorithm three‐stage are presented. simulation results verify validity obtained algorithms.
Abstract Accurate forecasting of traffic flow is crucial for intelligent control and guidance. It very challenging to forecast the due high non‐linearity, complexity dynamicity data. Most existing methods focus on designing complicated graph neural network architectures capture spatio‐temporal features data with help predefined graphs. However, exhibit a strong spatial dependency, which means that there are often complex correlations between nodes in road topology graph. Moreover, change...
An active area of study under the dual carbon target, which is based on automatic identification systems (AIS), emission inventory pollutants from ships. Data compression required because there currently so much data that it has become difficult to transmit, process, and store it. A trajectory simplification method considering ship sailing state acceleration rate change developed in this paper assure validity compressed used analysis. By carefully examining integral relationship between...
Network-centric operations demand an increasingly sophisticated level of interoperation and information fusion for escalating number throughput sensors human processes. The resulting complexity the systems being developed to face this environment render lower techniques alone simply insufficient ensure interoperability, as they fail consider subtle, but critical, aspects inherent in knowledge interchange. A fundamental mathematical theory high-level is needed address (1) representation...
Ship data obtained through the maritime sector will inevitably have missing values and outliers, which adversely affect subsequent study. Many existing methods for imputation cannot meet requirements of ship quality, especially in cases high rates. In this paper, a method based on generative adversarial networks (GANs) is proposed. The network (GAIN) improved using Wasserstein distance gradient penalty to handle values. Meanwhile, preprocessing process optimized by combining knowledge from...
Mixture distributions such as Gaussian mixture model (GMM) have been used in many applications for dynamic state estimation. These include robotics, image and acoustic processing, distributed tracking, multisensor data fusion. However, the recursive processing of incurs rapidly growing computational requirements. In particular, number components distribution grows exponentially when multiple them are combined. order to keep complexity tractable, it is necessary approximate a by reduced one...
Since UAV aerial images are usually captured by UAVs at high altitudes with oblique viewing angles, the amount of data is large, and spatial resolution changes greatly, so information on small targets easily lost during segmentation. Aiming above problems, this paper presents a semantic segmentation method for images, which introduces multi-scale feature extraction fusion module based encoding-decoding framework. By combining channel extraction, network can focus more certain layers regions...
The areas of three-dimensional (3D) underwater wireless sensor networks (UWSNs) have attracted significant attention recently due to their applications in detecting and observing phenomena that cannot be adequately observed by means two-dimensional UWSNs. However, designing routing protocols for UWSNs is a challenging task because path breakage occurs frequently uncertain node link failures. In this paper, we present Robust Routing Protocol (RRP) aims achieve robustness under harsh...
In this paper, we consider an adaptive node and power simultaneous scheduling (ANPSS) strategy for target tracking in distributed multiple radar systems. For all of the available nodes, with full resources allocation, minimizing estimation mean-square error (MSE) may exceed predetermined system performance goal cause unnecessary consumption. Therefore, driven resource allocation schemes systems are proposed. a predefined MSE threshold, total transmitted energy is minimized by optimally...
The multiple hypothesis tracker (MHT) is currently the preferred method for addressing data association problem in multitarget tracking (MTT) application. MHT seeks most likely global by enumerating all possible associations over time, which equal to calculating maximum a posteriori (MAP) estimate report data. Despite being well‐studied method, remains challenging mostly because of computational complexity association. In this paper, we describe an efficient solving using graphical model...
To prevent and reduce corporate financial risks, this paper builds a early-warning model for listed companies based on combination of SOM BP neural networks focusing short-term forecasting monitoring. Firstly, network is utilized to allow self-modification unit connection weights according the feature information input data enable weight vector distribution be similar sample data, thereby obtaining relatively optimal training samples among all samples. Then, monitoring created through...
Cloud computing offers resource-constrained users big-volume data storage and energy-consuming complicated computation. However, owing to the lack of full trust in cloud, cloud prefer privacy-preserving computation with correctness verification. cryptography-based schemes introduce high computational costs both its for verifiable privacy preservation, which makes it difficult support computations practice. Intel Software Guard Extensions (SGX) as a trusted execution environment is widely...
Abstract The optimization of phosphate (P) fertilizer application strategies to improve the utilization P in farmlands is urgently required. objective this study evaluate effect rate and position on maize ( Zea mays L.) utilization. A 2‐year field experiment was conducted with (0, 40, 80, 120 kg ha −1 ) (5 10 cm). Compared 5‐cm position, 10‐cm significantly increased available phosphorus (AP) content rhizosphere soil. Surprisingly, AP or remained same when amount reduced 80 at a distance 5...
The simplest hybrid Bayesian network is Conditional Linear Gaussian (CLG). It a model for which exact inference can be performed by the Junction Tree (JT) algorithm. However, traditional JT only provides first two moments hidden continuous variables. In general, complexity of algorithms exponential in size largest clique strongly triangulated graph that usually one including all discrete parent nodes connected component model. Furthermore, general nonlinear non-Gaussian model, it well-known...
In this paper, we propose an adaptive node selection strategy for target tracking in passive multiple radar systems, with the objective of minimizing number nodes task. Since signal parameters are random first take expectation over parameters, and derive a new Bayesian Cramer-Rao lower bound (BCRLB) as criterion. Then, formulate knapsack-based problem required accuracy constraint. This formulation can be solved optimally by exhaustive search algorithm, but high computational complexity. For...
The multiple hypothesis tracker (MHT) is a popular algorithm for solving multi-target tracking (MTT) problem in cluttered environment. It known as maximum posterior (MAP) estimator which enumerates all possible global hypotheses and dedicates to find the most likely solution based on received reports. However, its practical application often limited by complexity of data association step. This paper describes an efficient MHT "track-oriented" framework. proposed approach translates weight...