Based on genetic algorithm and neural network algorithm,the finite element analyses on the temperature fields and stress fields of multi-track laser cladding were carried out by using the ANSYS software.The results sh...Based on genetic algorithm and neural network algorithm,the finite element analyses on the temperature fields and stress fields of multi-track laser cladding were carried out by using the ANSYS software.The results show that,in the multi-track cladding process,the temperature field ellipse leans to the cladding formed,and the front cladding has preheating function on the following cladding.During cladding,the longitudinal stress is the largest,the lateral stress is the second,and the thickness direction stress is the smallest.The center of the cladding is in the tensile stress condition.The longitudinal tensile stress is higher than the lateral or thickness direction stress by several times,and the tensile stress achieves the maximum at the area of joint between the cladding and substrate.Therefore,it is inferred that transversal crack is the most main crack form in multi-track laser cladding.Moreover,the joint between cladding and substrate is the crack sensitive area,and this is consistent with the actual experiments.展开更多
To improve the efficiency and accuracy of single-event effect(SEE)research at the Heavy Ion Research Facility at Lanzhou,Hi’Beam-SEE must precisely localize the position at which each heavy ion hitting the integrated...To improve the efficiency and accuracy of single-event effect(SEE)research at the Heavy Ion Research Facility at Lanzhou,Hi’Beam-SEE must precisely localize the position at which each heavy ion hitting the integrated circuit(IC)causes SEE.In this study,we propose a fast multi-track location(FML)method based on deep learning to locate the position of each particle track with high speed and accuracy.FML can process a vast amount of data supplied by Hi’Beam-SEE online,revealing sensitive areas in real time.FML is a slot-based object-centric encoder-decoder structure in which each slot can learn the location information of each track in the image.To make the method more accurate for real data,we designed an algorithm to generate a simulated dataset with a distribution similar to that of the real data,which was then used to train the model.Extensive comparison experiments demonstrated that the FML method,which has the best performance on simulated datasets,has high accuracy on real datasets as well.In particular,FML can reach 238 fps and a standard error of 1.6237μm.This study discusses the design and performance of FML.展开更多
Bit-Pattemed Magnetic Recording(BPMR)is one of the emerging data storage technologies,which promises an Areal Density(AD)of about 4 Tb/in2.However,a major problem practically encountered in a BPMR system is Inter-Trac...Bit-Pattemed Magnetic Recording(BPMR)is one of the emerging data storage technologies,which promises an Areal Density(AD)of about 4 Tb/in2.However,a major problem practically encountered in a BPMR system is Inter-Track Interference(ITI)that can deteriorate the overall system performance,especially at high ADs.This paper proposes an iterative ITI cancellation method for an m-head m-track BPMR system,which uses m heads to read m adjacent tracks and decodes them simultaneously.To cancel the ITI,we subtract the weighted readback signals of adjacent tracks,acting as the ITI signals,from the readback signal of the target track,before passing the refined readback signal to a turbo decoder.Then,the decoded data will be employed to reconstruct the ITI signal for the next turbo iteration.Experimental results indicate that the proposed system performs better than the conventional system that uses one head to read one data track.Furthermore,we also find out that the proposed system is more robust to media noise and track misregistration than the conventional system.展开更多
为提高无人车障碍物检测跟踪的精度和稳定性,首先针对YOLO v5(You only look once version 5,YOLO v5)网络存在的语义信息和候选框信息丢失的问题,引入深度可分离空洞空间金字塔结构与目标框加权融合算法完成对网络的优化;其次针对单阶...为提高无人车障碍物检测跟踪的精度和稳定性,首先针对YOLO v5(You only look once version 5,YOLO v5)网络存在的语义信息和候选框信息丢失的问题,引入深度可分离空洞空间金字塔结构与目标框加权融合算法完成对网络的优化;其次针对单阶段障碍物点云聚类精度低的问题,设计一种考虑点云距离与外轮廓连续性的两阶段障碍物点云聚类方法并完成三维包围盒的建立;最后将注意力机制引入MobileNet使网络更加聚焦于目标对象特有的视觉特征,并综合利用视觉特征和三维点云信息共同构建关联性度量指标,提高匹配精度。利用KITTI数据集对构建的障碍物目标检测、跟踪与测速算法进行仿真测试,并搭建实车平台进行真实环境试验,验证所提算法的有效性和真实环境可迁移性。展开更多
基金Project(HIT(WH)200711)supported by the Study Fund of Harbin Institute of Technology at Weihai,China。
文摘Based on genetic algorithm and neural network algorithm,the finite element analyses on the temperature fields and stress fields of multi-track laser cladding were carried out by using the ANSYS software.The results show that,in the multi-track cladding process,the temperature field ellipse leans to the cladding formed,and the front cladding has preheating function on the following cladding.During cladding,the longitudinal stress is the largest,the lateral stress is the second,and the thickness direction stress is the smallest.The center of the cladding is in the tensile stress condition.The longitudinal tensile stress is higher than the lateral or thickness direction stress by several times,and the tensile stress achieves the maximum at the area of joint between the cladding and substrate.Therefore,it is inferred that transversal crack is the most main crack form in multi-track laser cladding.Moreover,the joint between cladding and substrate is the crack sensitive area,and this is consistent with the actual experiments.
基金supported by the National Natural Science Foundation of China (Nos.U2032209,11975292,12222512)the National Key Research and Development Program of China (2021YFA1601300)+2 种基金the CAS“Light of West China”Programthe CAS Pioneer Hundred Talent Programthe Guangdong Major Project of Basic and Applied Basic Research (No.2020B0301030008)。
文摘To improve the efficiency and accuracy of single-event effect(SEE)research at the Heavy Ion Research Facility at Lanzhou,Hi’Beam-SEE must precisely localize the position at which each heavy ion hitting the integrated circuit(IC)causes SEE.In this study,we propose a fast multi-track location(FML)method based on deep learning to locate the position of each particle track with high speed and accuracy.FML can process a vast amount of data supplied by Hi’Beam-SEE online,revealing sensitive areas in real time.FML is a slot-based object-centric encoder-decoder structure in which each slot can learn the location information of each track in the image.To make the method more accurate for real data,we designed an algorithm to generate a simulated dataset with a distribution similar to that of the real data,which was then used to train the model.Extensive comparison experiments demonstrated that the FML method,which has the best performance on simulated datasets,has high accuracy on real datasets as well.In particular,FML can reach 238 fps and a standard error of 1.6237μm.This study discusses the design and performance of FML.
基金This work was supported by Research and Development Institute,Nakhon Pathom Rajabhat University,Thailand,under Grant GP_6212.
文摘Bit-Pattemed Magnetic Recording(BPMR)is one of the emerging data storage technologies,which promises an Areal Density(AD)of about 4 Tb/in2.However,a major problem practically encountered in a BPMR system is Inter-Track Interference(ITI)that can deteriorate the overall system performance,especially at high ADs.This paper proposes an iterative ITI cancellation method for an m-head m-track BPMR system,which uses m heads to read m adjacent tracks and decodes them simultaneously.To cancel the ITI,we subtract the weighted readback signals of adjacent tracks,acting as the ITI signals,from the readback signal of the target track,before passing the refined readback signal to a turbo decoder.Then,the decoded data will be employed to reconstruct the ITI signal for the next turbo iteration.Experimental results indicate that the proposed system performs better than the conventional system that uses one head to read one data track.Furthermore,we also find out that the proposed system is more robust to media noise and track misregistration than the conventional system.
文摘为提高无人车障碍物检测跟踪的精度和稳定性,首先针对YOLO v5(You only look once version 5,YOLO v5)网络存在的语义信息和候选框信息丢失的问题,引入深度可分离空洞空间金字塔结构与目标框加权融合算法完成对网络的优化;其次针对单阶段障碍物点云聚类精度低的问题,设计一种考虑点云距离与外轮廓连续性的两阶段障碍物点云聚类方法并完成三维包围盒的建立;最后将注意力机制引入MobileNet使网络更加聚焦于目标对象特有的视觉特征,并综合利用视觉特征和三维点云信息共同构建关联性度量指标,提高匹配精度。利用KITTI数据集对构建的障碍物目标检测、跟踪与测速算法进行仿真测试,并搭建实车平台进行真实环境试验,验证所提算法的有效性和真实环境可迁移性。
文摘近年来,随着羊只养殖向大规模和精细化的方向发展,羊场对智能化管理的需求日益增加。因此,精准的个体识别和行为监测变得尤为重要,对多目标跟踪(Multiple object tracking, MOT)算法的准确性提出了更高要求。然而,现有的MOT算法在目标遮挡和动态场景下的性能仍不理想。本文提出两种跟踪线索:深度调制交并比(Depth modulated intersection over union, DIoU)和轨迹方向建模(Tracklet direction modeling, TDM),旨在补充交并比(Intersection over union, IoU)线索,提高多目标跟踪的精准度和鲁棒性。DIoU线索通过引入目标的深度信息改进了传统的IoU计算方法。TDM聚焦于目标的运动趋势,预测其未来的移动方向。本文将DIoU和TDM跟踪线索集成到BoT-SORT算法中,形成改进的多目标跟踪算法。在两个私有数据集上,改进算法相比基线方法,MOTA(Multiple object tracking accuracy)指标分别提高1.6、1.7个百分点,IDF1(Identification F1 score)指标分别提高1.9、1.0个百分点。结果显示,改进算法在复杂场景中的跟踪连续性和准确性显著提升。
