Junsuo Qu’s research while affiliated with Xi’an University of Posts and Telecommunications and other places

The human body is often occluded by a variety of obstacles in the monitoring system, so occluded person re-identification is still a long-standing challenge. Recent methods based on pose guidance or external semantic clues have improved the representation and related performance of features; there are still problems, such as weak model representation and unreliable semantic clues. To solve the above problems, we proposed a feature extraction network, named shared feature fusion with pose-guided and unsupervised semantic segmentation (SFPUS). This network will extract more discriminative features and reduce the occlusion noise on pedestrian matching. Firstly, the multibranch joint feature extraction module (MFE) is used to extract feature sets containing pose information and high-order semantic information. This module not only provides robust extraction capabilities but can also precisely segment occlusion and the body. Secondly, in order to obtain multiscale discriminant features, the multiscale correlation feature matching fusion module (MCF) is used to match the two feature sets, and the Pose–Semantic Fusion Loss is designed to calculate the similarity of the feature sets between different modes and fuse them into a feature set. Thirdly, to solve the problem of image occlusion, we use unsupervised cascade clustering to better prevent occlusion interference. Finally, performances of the proposed method and various existing methods are compared on the Occluded-Duke, Occluded-ReID, Market-1501 and Duke-MTMC datasets. The accuracy of Rank-1 reached 65.7%, 80.8%, 94.8% and 89.6%, respectively, and the mAP accuracy reached 58.8%, 72.5%, 91.8% and 80.1%. The experiment results demonstrate that our proposed SFPUS holds promising prospects and performs admirably compared with state-of-the-art methods.

In the 6G era, Space-Air-Ground Integrated Networks (SAGINs) can provide ubiquitous coverage for Internet of Things (IoT) devices. Multi-access Edge Computing (MEC) and blockchain are two enabling technologies, which can further enhance the services capabilities of SAGINs, where MEC demonstrates a notable capability in efficiently minimizing both task execution delays and system energy consumption, and blockchain can provide trust guarantee for task offloading and wireless data transmission among entities operated by different operators in SAGIN. In this paper, we present an MEC and blockchain enabled SAGIN architecture, which consists of two sub-systems. In the MEC sub-system, a satellite and multiple Unmanned Aerial Vehicles (UAVs) act as edge nodes to provide IoT devices with computing power. Moreover, the satellite serves as the block generator and the client, and the UAVs serve as consensus nodes of the blockchain sub-system. We intend to minimize the energy consumption within the network, which is achieved through the IoT devices’ task segmentation, the UAVs and satellite’s band-width allocation among their served IoT devices. And moreover, the computing power of UAVs and the satellite also allocated in task processing and blockchain consensus. Considering the high dynamics of the network, it is impossible to obtain real-time and accurate channel information, so we re-model this problem as a Markov decision process, and propose a low-complexity adaptive optimization algorithm based on Deep Deterministic Policy Gradient (DDPG). Our simulation results indicate that the proposed algorithm exhibits commendable performance in minimizing network energy consumption and DDPG agent’s accumulated reward maximization.

This paper presents the development of an advanced servo system for gluing operations, which integrates a dosing cylinder enhanced by reinforcement learning algorithms to optimize efficiency and accuracy. The system consists of an upper computer, a programmable logic controller (PLC), a main controller (featuring a core board and an IO board), and a dosing cylinder equipped with a servo motor and a glue gun. A comprehensive system model was developed by 1 analyzing the interactions among mechanical, hydraulic, and servo motor subsystems. The implementation of an enhanced Deep Deterministic Policy Gradient (DDPG) reinforcement learning algorithm, augmented with Model Predictive Control (MPC), enables precise control over voltage inputs and glue gun temperature adjustments. The MPC component predicts future states of the system based on the current state and actions, refining the action selection process by considering the long-term effects of actions. Rigorous simulation testing of the enhanced DDPG algorithm, alongside other algorithms, in configurations involving single and dual cylinder setups, validates its effectiveness, demonstrating significant enhancements in the automated control of the gluing process. The real-world experimental results indicate high positioning accuracy, consistent glue dispensing , and overall stability, making this system suitable for industrial applications requiring precise and reliable gluing.

Industrial large models are attracting significant attention for their roles in improving industrial production efficiency and product quality. This paper categorises and reviews current research on industrial large models in three main areas: pre-training, fine-tuning, and Retrieval-Augmented Generation (RAG). It also introduces a generic platform for industrial large models, including a model for interaction between industrial large and small models. Furthermore, it specifies the application areas of large industrial models within product lifecycle management, and discusses the challenges encountered during their development.

Vehicular mobile edge computing (vMEC) and non-orthogonal multiple access (NOMA) have emerged as promising technologies for enabling low-latency and high-throughput applications in vehicular networks. In this paper, we propose a novel multi-agent deep deterministic policy gradient (MADDPG) approach for resource allocation in NOMA-based vMEC systems. Our approach leverages deep reinforcement learning (DRL) to enable vehicles to offload computation-intensive tasks to nearby edge servers, optimizing resource allocation decisions while ensuring low-latency communication. We introduce an attention mechanism within the MADDPG model to dynamically focus on relevant information from the input state and joint actions, enhancing the model’s predictive accuracy. Additionally, we propose an attention-based experience replay method to expedite network convergence. The simulation results highlight the effectiveness of multi-agent reinforcement learning (MARL) algorithms, such as MADDPG with attention, in achieving better convergence and performance in various scenarios. The influence of different model parameters, such as input data volumes, task load levels, and resource configurations, on optimization results is also evident. The decision making processes of agents are dynamic and depend on factors specific to the task and environment.

This paper presents an in-depth study of the application of Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithms with an exploratory strategy for duty cycle scheduling (DCS) in the wireless sensor networks (WSNs). The focus is on optimizing the performance of sensor nodes in terms of energy efficiency and event detection rates under varying environmental conditions. Through a series of simulations, we investigate the impact of key parameters such as the sensor specificity constant α and the Poisson rate of events on the learning and operational efficacy of sensor nodes. Our results demonstrate that the MADDPG algorithm with an exploratory strategy outperforms traditional reinforcement learning algorithms, particularly in environments characterized by high event rates and the need for precise energy management. The exploratory strategy enables a more effective balance between exploration and exploitation, leading to improved policy learning and adaptation in dynamic and uncertain environments. Furthermore, we explore the sensitivity of different algorithms to the tuning of the sensor specificity constant α, revealing that lower values generally yield better performance by reducing energy consumption without significantly compromising event detection. The study also examines the algorithms' robustness against the variability introduced by different event Poisson rates, emphasizing the importance of algorithm selection and parameter tuning in practical WSN applications. The insights gained from this research provide valuable guidelines for the deployment of sensor networks in real-world scenarios, where the trade-off between energy consumption and event detection is critical. Our findings suggest that the integration of exploratory strategies in MADDPG algorithms can significantly enhance the performance and reliability of sensor nodes in WSNs.

In scenarios where road obstacles complicate feature extraction, designing a lightweight CNN model with minimal parameters and flops while maintaining competitive segmentation accuracy poses one of the most challenging research tasks in remote sensing imaging. Finding the optimal balance between segmentation performance and computational efficiency is crucial. We introduces a novel method for global road feature extraction by strategically employing the Light Ghost Basic block to develop a tiny Network (TG-LinkNet). A Multi-Scale Feature Fusion (MSFF) module, which combines the Parallel Channel Position Attention Mechanism (PCPAM) to deliver accurate road structure information, further supports the goal. We present a solution to the issue of feature fusion information retrieval-induced excessive redundant noise, which might cause serious interference. Furthermore, to efficiently extract edge features and capture long-distance reliance on global features, we create a Global Context Feature Extraction (GCFE) module. Ultimately resulting in the Lightweight Global Road Feature Extraction Network (LGRF-Net). To facilitate efficient training, we implement a 1:2 weight design within our deep supervision technique, termed hybrid loss (WCE-Dice). Extensive experiments were conducted on the DeepGlobe (1024 × 1024, 512 × 512) and SpaceNet road datasets. This demonstrates that our network with smaller parameters and flops compared to other road-based semantic segmentation methods.

Due to obstructions such as trees and buildings, single-modal satellite or aerial images are insufficient for continuous high-precision representation of road features. To address this problem, this paper proposes a Lightweight Cross-Modal Information Interaction for Road Feature Extraction (LCIRE-Net) from High-Resolution Remote Sensing Images (HRSIs) and GPS Trajectory/LiDAR images. We design two parallel encoders for modality feature learning, using pairs of multi-modal information as inputs to the encoders. By designing a Cross-Modal Information Dynamic Interaction (CMIDI) mechanism, thresholds are used to decide whether to supplement redundant information from another modality, solving the issue of ineffective fusion calculations due to minor differences in multi-modal feedback. A Multi-modal Feature Fusion Module (MFFM) is proposed after the encoder output to achieve effective dual-modal fusion while addressing the interference of redundant noise generated during extraction. Subsequently, we present the Feature Refinement and Enhancement Module (FREM), which successfully captures edge features of the image using the receptive field of dilated convolution kernels. Additionally, in terms of lightweight design, we employ a novel SOTA method on D-LinkNet by replacing the original residual blocks with an enhanced Ghost Basic Block. Extensive experiments are conducted on the BJRoad, Porto, and TLCGIS datasets, demonstrating that our network, with smaller parameters and FLOPs, outperforms other road-based semantic segmentation methods.

In this paper, we discuss the issue of bearing fault diagnosis in impulsive environments. Such impulsive signals have significant spike impulse characteristics and show the obvious non-Gaussian property. Compare to the cyclic impulsive signals generated by bearing local damage, the impulsive components can be considered to be a special kind of noise, namely impulsive noise. Unfortunately, the impulsive nature of the noise often leads to significant degradation of the performance of the signal processing techniques based on the Gaussian model. To overcome this issue, an impulsive noise filtering method based on alpha-stable distribution (α-stable filter for short) is used in the work. The α-stable filter is capable of attenuating the energy of impulsive components to a large extent, thereby establishing a prerequisite for subsequent fault feature extraction. However, the performance of the α-stable filter is greatly dependent on the appropriate selection of the parameter of the filter order. Thus, to avoid the blind selection of the filter order in the α-stable filter and further improve its performance, an optimized α-stable filter is proposed in this paper. Firstly, the classical particle swarm optimization (PSO) is used to combine with the α-stable filter for selecting an optimal parameter. Secondly, a new index is constructed by combining the Gini coefficient with envelope spectrum kurtosis (ESK). The two indicators are robust against outliers, making them suitable for impulsive environments. The optimization objective function of the PSO is enhanced using the new index. We apply the optimized α-stable filter is applied to both simulated and real signals. The obtained results demonstrate that the filtering method is effective in canceling impulsive noise and enhances the ability to bearing fault detection.