赵卫东，工学博士，研究员，博士生导师。现任bat365在线官方网站企业数字化技术教育部工程研究中心常务副主任、上海工业视觉感知与智能计算工程技术研究中心主任、全国工业自动化系统与集成标准化技术委员会工业数据分技术委员会副主任委员、上海市系统仿真协会副理事长、2030国家重大专项专家、国家重点研发计划《智能工业软件》专家、上海市工业软件重点研发专项专家组召集人。
研究方向：人工智能、工业软件与智能制造。聚焦多模态具身感知、空间智能、视觉—语言—动作模型、机器人自主控制与轨迹优化，以及复杂制造系统智能排产、多智能体协同决策和数字孪生优化。
科研成果：累计发表高水平学术论文近百篇，授权发明专利40 余项；带领和支持团队教师作为项目负责人获批国家重点研发计划7项，2030国家重大专项课题多项。成果荣获上海市科技进步一等奖1 项、二等奖 2 项、三等奖 3 项。
科研论文：
[1]Hu, Lingyun and Liu, Xianhui and Wang, Shengyi and Jiang, Siqi and Shen,Jiaji and Zhao, Weidong.A dual-enhancement framework for constraint-aware job shop schedulingwith large language models[J].JOURNAL OF MANUFACTURING SYSTEMS.2026.（中科院一区，JCR一区）
[2]Gao, Gong and Zhao, Weidong and Liu, Xianhui and Jia, Ning.Improving policy exploitation in online reinforcement learning withinstant retrospect action[J].NEURAL NETWORKS,2026.（中科院二区TOP，JCR一区）
[3]Wei, Mingyue and Zhao, Weidong and Jia, Ning and Liu, Xianhui and Xu,Zhen.MSDD-Net: A lightweight multi-scale defect detection network forindustrial lead frame inspection[J].ALEXANDRIA ENGINEERING JOURNAL. 2026. （中科院二区TOP，JCR一区）
[4]Liu, Xianhui and Yang, Yangyang and Zhu, Chenlin and Hu, Yihan and Zhao,Weidong.LE OS: A lightweight edge operating system for industrial internet ofthings under resource constraints[J].FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE. 2026.（中科院二区TOP，JCR一区）
[5]Yu, Deshui and Liu, Xianhui and Ning, Jia and Wang, Shengyi and Zhu,Chenglin and Zhao, Weidong.Deep Reinforcement Learning-Based AI Task Offloading inResource-Constrained IIoT Computing Environments[J].IEEE INTERNET OF THINGS JOURNAL. 2025.（中科院一区TOP，JCR一区）
[6]Zhang, Haotian and Hou, Wenlong and Jia, Ning and Zhao, Weidong.An Improved Neighborhood-Based Mismatch Removal Algorithm for ConsumerDevice[J].IEEE TRANSACTIONS ON CONSUMER ELECTRONICS. 2025.（中科院二区TOP，JCR一区）
[7]Wang, Jiayin and Yao, Mingfeng and Wei, Yanran and Guo, Xiaoyu andZheng, Ayong and Zhao, Weidong.Image-to-Force Estimation for Soft Tissue Interaction inRobotic-Assisted Surgery Using Structured Light[J].IEEE ROBOTICS AND AUTOMATION LETTERS. 2025.（中科院二区，JCR一区）
[8]Hou, Wenlong and Zhao, Weidong and Jia, Ning and Liu, Xianhui.Low-resource knowledge graph completion based on knowledge distillationdriven by large language models[J].APPLIED SOFT COMPUTING. 2025.（中科院二区TOP，JCR一区）
[9]Yang, Jiaxiong and Jia, Ning and Liu, Xianhui and Fan, Rui and Sun,Yougang and Zhao, Weidong.Zone-YOLO: Vision-Language Object Detection Using Zone Prompt[J].IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS. 2025.（中科院一区TOP，JCR一区）
[10]Xu, Zhen and Zhao, Weidong and Jia, Ning and Liu, Xianhui and Wei,Mingyue.A Cascaded Network for Surface Defect Detection on Lead Frames inProduction Lines[J].IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT. 2024.（中科院二区TOP，JCR一区）
[11]Hou, Wenlong and Wu, Wenda and Liu, Xianhui and Zhao, Weidong.Document-level relation extraction with multi-semantic knowledgeInteraction[J].INFORMATION SCIENCES. 2024.（中科院二区TOP，JCR一区）
[12]Chen, Guojie and Zhao, Weidong and Liu, Xianhui and Wei, Mingyue andGao, Gong.Multi-Domain Merging Adaptation for Container Rehandling ProbabilityPrediction[J].IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS. 2024.（中科院一区TOP，JCR一区）
[13]Xie, Ziqi and Zhao, Weidong and Liu, Xianhui and Xu, Zhen and Zhao, Jianand Gao, Gong.RectanglingGAN: Deep rectangling for stitched image via image inpainting[J].KNOWLEDGE-BASED SYSTEMS. 2024.（中科院一区TOP，JCR一区）
授权专利：
[1] bat365在线官方网站,福建金岳慧创智能科技有限公司. 一种基于微调大语言模型的飞机总装信息关联规则挖掘方法、设备及介质: 202510685297.8[P]. 2025-08-12.
[2]bat365在线官方网站. 基于深度学习的微小目标缺陷识别模型训练方法: 202510948990.X[P]. 2025-09-02.
[3]bat365在线官方网站. 一种基于奖励累积深度强化学习的Kubernetes微服务最优部署方法、设备及介质: 202511255357.9[P]. 2025-12-23.
[4] bat365在线官方网站. 基于深度一致性多模态数据联合增强方法及系统: 202411320613.3[P]. 2025-10-31.
[5] bat365在线官方网站. 用于港口资源配置的边界感知数据增强方法: 202411943768.2[P]. 2025-10-10.
