In this paper, an anomaly monitoring method based on the fusion of particle filtering (PF) and ultra-wideband (UWB) technologies is proposed to address the challenge of real-time monitoring of anomalous states—such as leakage, illicit transfer, and source-container separation—of radioactive sources in nuclear materials repositories. Firstly, radiation field data is acquired within the repository through a fixed NaI(Tl) scintillator detector array, and the equivalent position of the radioactive source is inverted using a particle filtering algorithm. Concurrently, high-precision two-dimensional tracking of the nuclear material container, equipped with a UWB tag, is conducted using the time difference of arrival (TDOA) algorithm. Then, by establishing a real-time correlation model between the UWB-derived physical coordinates and the PF-derived radiation coordinates, the monitoring dimension is innovatively expanded from basic position tracking to comprehensive state discrimination. Finally, experimental results demonstrate that the proposed method not only overcomes the interference of environmental shielding and multipath effects but also effectively distinguishes anomalous events, including normal material movement, radioactive source leakage, and source-container separation, by analyzing the deviation and dispersion between the two coordinate systems. Ultimately, this research provides a highly robust, multi-modal proactive early-warning solution for the security of nuclear facilities.Experimental results demonstrate that under normal conditions, the fusion method reduces the trajectory RMSE to 0.35m. During anomaly detection, the system successfully identifies source-container separation with a spatial deviation exceeding 6.02m​

Accurate perception of three-dimensional radiation environments is essential for nuclear accident response and nuclear facility decommissioning. This paper presents a mobile robot-based method for 3D radiation field reconstruction. Unlike conventional interpolation approaches, the proposed approach establishes a physics-guided framework in which source localization supports radiation field mapping. An autonomous patrol-to-source-seeking switching strategy is first designed using a mechanical collimator. Angle-constrained particle filtering and maximum likelihood estimation are then employed to rapidly identify source parameters. These estimates provide accurate physical priors for subsequent field reconstruction. A physics-informed heteroscedastic Gaussian process model, termed PI-HGPR, is further introduced for radiation field inference. The model captures fluctuations in high-radiation regions by incorporating heteroscedastic noise. It also incorporates a physics-based attenuation kernel to reduce oversmoothing under sparse sampling. Finally, the reconstructed continuous radiation field is fused with a 3D geometric map to generate a radiation point cloud. Experimental results demonstrate that the proposed method maintains high reconstruction accuracy under sparse-sample conditions, achieving an RMSE below 7%, and provides a reliable data foundation for unmanned nuclear emergency exploration.

Locating multiple unknown radioactive targets in high-uncertainty environments is a critical challenge complicated by the severe "masking effect" of strong sources and "range ambiguity" from isotropic signal fields. In this paper, the application in engineering is the autonomous monitoring and physical protection of nuclear materials using a heterogeneous multi-robot swarm deployed in unknown, unstructured environments. The contribution in Artificial Intelligence (AI) is the development of a spatial-probabilistic decoupled active sensing framework for distributed multi-agent systems. To overcome the cognitive deadlocks inherent in traditional search strategies, the proposed AI architecture first employs a spatial-level Voronoi topological mechanism to structurally partition the globally complex multi-modal field, thereby neutralizing dominant source interference. Subsequently, each robot maintains a distributed Probabilistic Knowledge Base via a sequential particle filter. By mathematically establishing an analytical equivalence between orthogonal angle of arrival prediction variance and mutual information gain, the high-dimensional information-theoretic optimization is reduced to an efficient O(Np) variance calculation, which serves as the real-time AI perception decision rule. Upon target confirmation, a sequential masking and residual check mechanism dynamically peels off the confirmed signals, driving the distributed reasoning chain to sequentially uncover heavily concealed anomalies. Extensive simulations and real-world hardware-in-the-loop physical deployments validate the framework's efficacy. Compared with conventional single-modal algorithms, the terminal localization root mean square error is significantly reduced by 56.2%, achieving a 95% confidence level within a 1.5-meter radius, while the total swarm trajectory cost is halved.

In this paper, an intelligent mobile distributed measurement system is proposed to locate multiple unknown radioactive sources in complex environments. To overcome traditional measurement bottlenecks such as multi-source interference, particle degeneracy in zero-signal areas, and epistemic uncertainty, the system integrates Dynamic Voronoi Partitioning (DVP) with a novel dual-layer data fusion mechanism. Firstly, a DVP mechanism dynamically allocates exclusive sensing sub-regions based on real-time node topologies, ensuring concurrent and conflict-free data acquisition. Secondly, a dual-layer map-driven architecture—incorporating a Probability of Detection (POD) map for hypothesis erasing and a Gaussian situational map for continuous intensity field reconstruction—is designed to guide mobile nodes toward highly uncertain areas. Crucially, the system is explicitly initialized with rough priors containing severe position deviations and false alarms. By rigorously fusing real-time counts per second (CPS) measurements, the nodes autonomously filter out initial errors and reconstruct a global radiation field intensity map in real time. Simulation results in a 5-node, 1-source (5V1) scenario confirm that the proposed framework compresses localization measurement errors by 66.7\% (from 0.30 m to 0.10 m) compared to standard non-cooperative tracking. Furthermore, rigorous real-world physical experiments involving three robots and two sources (3V2) validate the structural superiority of the framework. Compared to an uncoordinated three-robot baseline swarm, the proposed cooperative measurement approach eliminates spatial aggregation, drastically reducing redundant total travel distance from 56.8 m to 21.5 m and compressing search time from 35 s to 19 s, while maintaining mean localization errors of 0.20 m for both sources.

Conducting source search and localization in an unknown radiation environment using unmanned ground vehicles (UGVs) is one of the critical challenges in nuclear safety and emergency response. In multi-source contamination scenarios, sensors can only measure the coupled field intensity, which complicates source separation and localization. This paper proposes an angle suppression multi-layer parallel particle filtering (ASMP) framework for multi-source term estimation. Initially, the algorithm utilizes the fusion of particle filtering with azimuthal angle information to partition parallel particle sets for estimating the positions of radiation sources. Subsequently, an improved Gaussian Process Regression (IGPR) is employed to predict the radiation field and provide prior knowledge for source term estimation (STE). Finally, an adaptive search strategy is integrated into the UGV to enhance the flexibility of the search process. Both simulation and real-world experimental results demonstrate that the proposed method effectively executes multi-source search tasks, achieving high localization accuracy, reduced localization uncertainty, and robustness under complex radiation conditions.

To enhance the precision of locating unidentified radiation sources with a coded aperture camera, this study introduces a refined Maximum Likelihood Expectation Maximization (MLEM) algorithm. This algorithm incorporates the dichotomy method to estimate the distance to the radiation source, followed by a superimposition of the calculated system response matrix to predict the source's position within the imaging area. Initially, a system response matrix table is created, listing potential distances from the radiation source to the camera in ascending order. The process involves assessing the mean square error (MSE) and the full width at half maximum (FWHM) of the current image to determine if the hypothesized distance surpasses the actual distance. Following this, a binary search is executed on the matrix table to ascertain the closest lower and upper bounds to the true distance. Subsequently, the system matrices for these boundary distances are aggregated to derive an approximate response matrix. This matrix is then employed in the MLEM algorithm to estimate the radiation source's position. Experimental outcomes indicate that the enhanced algorithm provides superior localization accuracy and a definable distance range (nearest lower and upper bounds) between the radiation source and the camera compared to the conventional MLEM algorithm, thus better facilitating the analysis of hotspots.The proposed method achieves satisfactory reconstruction for a single radiation source and for multiple sources located at the same detector distance; however, its performance in reconstructing multiple sources at differing distances remains to be improved.

This study addresses the situational awareness requirements for unmanned surface vehicles(USVs)operating in nearshore environments by proposing a dual solid-state LiDAR-based method for detecting water surface targets.The proposed method achieves high-resolution detection at a relatively low cost.First,a joint calibration model for the two solid-state LiDARs is developed utilizing the least squares method to determine the extrinsic parameters of the LiDAR scanning system.This calibration broadens the horizontal field of view while maintaining a 144-line scanning resolution.Second,a joint calibration between the LiDAR scanning system and the inertial measurement unit is performed,which reduces point cloud distortion through pose calibration and corrects scanning errors caused by the vessel's motion.In the detection algorithm,adaptive kernel convolution and hierarchical attention mechanisms are integratedinto the CenterPoint network,resulting in significant improvements in both inference speed and detection accuracy.A dedicated dataset for nearshore situational awareness is created,and experimental validation is conducted.Results show that the proposed algorithm improves overall average precision for bird's eye view(BEV AP)and 3D AP by 6.14 percentage points and 4.66 percentage points,respectively,and increases inference speed by more than 40%as compared with the baseline algorithm.

The deployment of mobile robots to conduct search for radioactive sources plays a crucial role in preventing radiation pollution and safeguarding public health. And it is a significant challenge to accurately locate radioactive sources in unknown environments. Herein, a particle filter based on mobile robot search method is proposed to localize radiation sources. By applying an angle constraint technique, the complexity of this algorithm is reduced, the high accuracy of source estimation is maintained, and the range of search is expanded. Additionally, a particle diffusion technology is applied to address particle loss that occurs during filtering. Furthermore, an adaptive gradient descent strategy is adopted to enhance the search efficiency. Experimental results demonstrate that this method achieves a success rate of over 95 % within a rectangular area (e.g., 40m by 40m), outperforming traditional methods, and it could perform effectively in dual radioactive source search tasks.

未知放射源搜寻定位是核安检、核应急领域的重要研究课题。为提高寻源效率、适应多源环境探测,提出了一种融合到达角(AOA)的粒子滤波寻源方法。首先,构建了自主定位与AOA感知相结合的硬件平台,在搜寻过程中给探测器引入了位置和角度信息;其次,在粒子滤波基础上考虑AOA信息,动态收缩放射源搜寻区域,进而提高搜寻效率;最后,在自主寻源路径规划中采用AOA引导的机器人姿态调整,增强机器人寻源的灵活度。实验证明此方法可正确有效工作,放射源校验测试证明该方法也适用于多源搜寻。

As radioactive sources are increasingly used in scientific research,industry and other fields,radioactive sources are often lost due to improper management and control.Since gamma rays have higher penetrating power and longer action distance than a particles and β rays,they can affect a larger area and have radiation effects on various substances in the environment.Therefore,the search and positioning of radioactive sources has always been the focus of radioactive source search.Although gamma detection equipment can achieve relatively precise positioning of a single radioactive source,when there are multiple radioactive sources in the radiation environment,radioactive sources with weaker energy can easily be ignored.In order to enable the gamma camera to accurately locate multiple radioactive sources in the unknown radiation field at the same time,this paper proposed a multi-distance system matrix superposition gamma image reconstruction algorithm based on the maximum likelihood expectation maximization(MLEM)algorithm.First,Monte Carlo simulation software was used to build a gamma camera simulation,and the impact of different encoding patterns on detection efficiency and the impact of radioactive sources at different distances on detection performance were analyzed.In order to be as close to the actual situation as possible,the photoelectric effect,Compton scattering,electron pair effect and decay of the radioactive source were considered in subsequent simulation experiments,and the simulation structure is consistent with the gamma camera structure.Then,simulation was used to obtain projection data,and different distance system matrices were calculated based on the impact of radioactive source distance on detector collection efficiency.Besides,in order to simultaneously located multiple radioactive sources within different distance ranges,a new system matrix was formed by superimposing multiple system matrices with different distances,and then calculated with the projection data,and through continuous iteration,a higher quality reconstructed image and the position information of multiple radioactive sources were obtained.Finally,in order to verify the feasibility of the algorithm,the algorithm was used for actual verification of 60Co and 137Cs radioactive sources in a real environment,accurately reconstructed the position information of the two radioactive sources.Experimental results show that this method can quickly locate radioactive sources.It has high convergence when using a single source,and it can accurately locate multiple radioactive sources when using multiple sources.It has a longer imaging distance and higher positioning accuracy than related decoding algorithms.This method is practical and feasible in practical applications.

This letter presents a reconfigurable leader-follower formation approach using relative pose estimation to steer a robot team to maintain a geometry pattern and pass through narrow spaces in unknown environments. The relative pose is estimated using a nonlinear optimization process and the sliding window approach by minimizing the residual error of short-term peer-to-peer UWB (Ultra-Wideband) ranging and odometry. We enhance the estimation accuracy using pose graph optimization by additionally considering odometry as constraints. To pass through narrow areas in unknown environments, we present a pattern reconfiguration mechanism to determine the formation styles (geometry pattern or linear passing pattern) based on environment traversability. A tracking controller is designed to steer the robots to achieve the desired formation style. The effectiveness and performance of the proposed approach are verified by experiments in an indoor environment with ground nonholonomic robots.

In this paper, it is proposed to locate multiple unknown radioactive sources within a certain time limit through particle filtering and Voronoi partitioning. Firstly, with each robot as a Voronoi centroid, the entire area is partitioned. Then, the robots conduct source search concurrently through particle filtering. When all the robots complete the process of one-particle filtering, the iteration ends and the next one begins until the search for the radioactive source is terminated. Finally, experiment is conducted to demonstrate the efficiency and accuracy of the proposed method.

The intensive application of radioactive sources in various fields has led to a rise in incidents involving their loss. Focusing on the rapid and safe detection of radiation environments and the localization of radioactive sources, this paper integrates the design of detection robots, radiation localization techniques, and path tracking to create a radioactive source detection system based on autonomous mobile robots. The design of the detection robot considers the unique challenges posed by radiation environments. In terms of localization, an adaptive radioactive source localization algorithm based on the Bayesian model, incorporating angle of arrival (AOA), is proposed to estimate the location of radioactive sources. Path planning is optimized using the Pure Pursuit algorithm to enhance tracking accuracy. Finally, on a mobile robot platform, an unknown radioactive source search experiment is conducted, validating the effectiveness of the system and vigorously advancing the field of radioactive source detection.

Wheel odometry often does not perform as well as expected on complex surfaces, uneven surfaces, and smooth ground. At the same time, the traditional laser scan matching method does not always correctly correlate the relationship between point clouds, and thus is likely to have abnormal point-to-point correlation, which leads to bad localization accuracy. To solve this problem, a laser scanning matching method based on directional endpoints is proposed. First, we extract straight-line endpoints from the environment as the feature points, and then use the feature matching between the endpoints to obtain the relative pose relationships of the mobile robots in adjacent moments. The directional endpoints are used to eliminate the mismatched feature points to further improve the matching accuracy. Hence, the iterative closest point method is used to further optimize the matching results of the directional endpoints to obtain a better localization result of the mobile robot. The experiment results show that the method achieves an average localization error and an average angle error of 0. 12 m and 1. 18 degrees, respectively, in an indoor environment of 7 m x 7 m, which is superior in accuracy compared with the traditional laser scan matching algorithm.

In response to the demand for customs detection and tracking of individuals carrying radioactive sources, this paper proposes a method that integrates radiographic imaging information for identifying and tracking carriers. Initially, a portable source personnel visual tracking system equipped with a gamma camera is constructed to capture the position of radioactive sources and images of moving personnel within the security inspection area. Subsequently, a method is developed to link personnel targets identified by both the gamma camera and surveillance cameras, using Euclidean distance and intersecting areas to pinpoint carriers. Furthermore, leveraging the improved RepVGG reparameterization structure and residual networks, we accelerate Re-Identification (RE-ID) inference and propose a Deep Feature Tracking (DFT) method. This method tracks personnel by comparing feature similarities in adjacent frames. Experimental results validate the effectiveness of this method in identifying suspicious individuals carrying radioactive sources and in mapping the walking trajectories of these carriers using pre-deployed surveillance cameras.

针对复杂环境下移动机器人可靠作业需求,结合机器人自主导航、网络测量、通信组网等技术,设计可感知作业环境通信质量的中继通信机器人,为机器人调度、任务规划提供依据.主要内容包括:设计移动中继机器人,增强作业机器人与控制台之间通信质量;设计通信质量感知的地图构建方法,实现机器人作业环境通信质量的可视化;设计通信中断自恢复机制,实现通信中断情况下机器人的自主返回与通信连接恢复.

安全帽佩戴识别是一种分类少的目标检测任务，使用现有精度较高的大型深度学习网络模型来识别安全帽佩戴，存在参数冗余问题且计算较大不利于部署在计算量有限的嵌入式设备中以适应实际的工地环境。针对以上问题，提出了一种适合部署在嵌入式设备中的轻量化网络模型YOLO-Ghost-BiFPNs3。在YOLOv4的基础上，基于Ghost模块重构形成新的网络结构幷对网络的深度和宽度进行裁剪；设计一种基于通道加权相加的轻量化模块BiFPNs3来替换原来计算量较大的FPN+PAN的结构；采用更容易量化的h-Swish激活函数；在Safety-Helmet-Wearing-Dataset数据集上进行实验，在测试集上，mAP@0.5为91.1%，相较于YOLOv4精度仅损失一个百分点，比轻量化网络模型YOLOv4-Tiny精度高26个百分点。参数量为原来YOLOv4的3%，计算量仅为原来YOLOv4的5.8%。

Aiming at the problem of multi-objective optimization conflict between makespan and average resource utilization in cloud computing task scheduling, this paper proposed a cloud computing task scheduling strategy based on multi-strategy improved harris hawk optimization algorithm. Firstly, the fusion strategy of Tent chaotic sequence and opposition-based learning is introduced to initialize the population to enhance the diversity of the population and the quality of the initial solution, thereby improving the optimization efficiency of the population; Secondly, an energy adaptive exponential decay factor is employed as a control parameter for global exploration and local exploitation to improve the accuracy and convergence speed of the algorithm; Finally, apply the elite representative exploration mechanism of the grey wolf optimization (GWO) algorithm to the global search stage of the population to improve the global search ability of the algorithm and avoid falling into a local optimal state. The simulation results show that for task scheduling with different distributions and scales, the proposed algorithm is superior to the comparison algorithm in terms of reducing makespan and increasing the average resource utilization rate, and has obvious advantages in algorithm convergence and convergence accuracy.

可靠定位是机器人完成导航和路径规划的前提,机器人通过多个超宽带(ultra-wideband,UWB)基站的测距信息实现定位,但基站数量不足时定位精度受限.针对这一问题,提出融合超宽带距离和方位的移动机器人定位方法.根据方位标准差区分信号来自基站前方(视场)或背后(非视场),消除方位的前后奇异性.在此基础上,利用UWB距离和方位测量值构建约束函数,通过图优化算法融合里程计和UWB测量数据实现全局位姿优化.实验结果表明,该方法在13 m×6 m的室内环境中,移动机器人无规则运动能够达到0.093 m的定位精度,比传统的基于测距UWB和里程计融合方法定位性能提升了 46％,且具有较强的鲁棒性.