施巍松老师组20-21年工作

Qingyang Zhang, Hong Zhong, Weisong Shi, Lu Liu: A trusted and collaborative framework for deep learning in IoT. Comput. Networks 193: 108055 (2021)

物联网应用程序包含大量用户隐私,将数据推送到其他应用程序可能会导致隐私泄露。基于可信执行环境技术,我们提出了一种框架,该框架可以在计算安全性和传输安全性方面为未来支持人工智能的物联网提供可信协作,在这种环境中可以对数据进行处理,并提出了两种方法(分别针对TEE完全可信和TEE部分可信)来保证数据传输的安全性。

框架图:
在这里插入图片描述
a TEE-enabled pure AI model inference application will be launched as multiple instances, and connected by a pipeline, which consists of several message queues created by the Task Manager module. Therefore, the size of model inferred by one TEE instance could be limited to avoid out of memory on TEE hardware. (会设置队列管理,避免超出TEE的内存限制,但这有点太straightforward了吧,感觉和引言中的挑战描述:In this case, the memory overhead will much larger than the size of protected memory in TEE, resulting in high latency overhead caused by frequent memory exchange between protected memory and normal memory. 不相匹配)

针对用户是否信任TEE允许TEE进行解密,设计了两种方案。前者使用对称加密,每次TEE加密解密。后者使用非对称加密,但非对称加密开销大,所以只用来加密对称加密的密钥。其核心是想让在一个公钥下加密的密文转换为在另一个公钥下加密的密文,同时不会泄漏底层消息或私钥。具体做法貌似是把key拆分成了两部分,分别进行处理。
在这里插入图片描述

Qingyang Zhang, Hong Zhong, Jie Cui, Lingmei Ren, Weisong Shi: AC4AV: A Flexible and Dynamic Access Control Framework for Connected and Autonomous Vehicles. IEEE Internet Things J. 8(3): 1946-1958 (2021)

面向智能网联车场景,其访问控制机制需满足以下三点:1. 未知的应用需要新的访问控制,需要具有可扩展性;2. 细粒度控制,如方向盘的控制数据只能由某些应用所控,其他应用不可见;3. 需要根据场景和系统状态动态变化;4. 需要支持不同的访问控制策略。

因此,本文提出了一种新颖灵活的访问控制框架AC4AV,该框架旨在支持各种访问控制模型,并为动态调整访问控制模型和开发定制的访问控制模型提供api,从而支持社区对CAV的访问控制研究。此外,本文还提出了一种数据抽象方法,可以清晰地识别CAV中的数据、应用程序和访问操作,从而可以方便地在访问控制策略中配置每个数据和应用程序的权限。

Liangkai Liu, Sidi Lu, Ren Zhong, Baofu Wu, Yongtao Yao, Qingyang Zhang, Weisong Shi: Computing Systems for Autonomous Driving: State of the Art and Challenges. IEEE Internet Things J. 8(8): 6469-6486 (2021)

七大性能评估指标:Accuracy,Timeliness,Power,Cost,Reliability,Privacy,Security
九大关键技术:

  1. Sensors:Camera(易受环境因素影响,数据体积大),Radar(价格贵),LiDAR(更为准确但价格更贵),Ultrasonic Sensor (短距离),GPS/GNSS/IMU (精度差,错误随时间累积)
    在这里插入图片描述
  2. Data Source:Data characteristics (实时数据和历史数据); Dataset and Benchmark;Labeling
  3. Autonomous Driving Applications:目标检测;车道检测;定位和地图;预测和规划;车辆控制
  4. Computation Hardware
  5. Storage:确保传感器收集正确的数据,它被立即处理,安全地存储,并转移到链中的其他技术,例如路边单元(RSU)、云数据中心,甚至第三方用户。层级存储和工作流。
  6. Real-Time Operating Systems
  7. Middleware Systems
  8. Vehicular Communication
  9. Security and Privacy:感知安全(jamming attacks and spoofing attacks);传输安全;数据安全(传输和存储过程中);控制安全(replay attack,relay attack);隐私(采集到行人的人脸数据、车辆牌照信息等如何存储,地理位置相关的服务)
    十二大技术挑战:
  10. Artificial intelligence for AVs:安全问题的标准;可扩展性(PB级数据;精确的标注);完整的测试(耗时很长)
  11. Multi-sensors Data Synchronization(多个传感设备之间的采集数据频率不同,如何在存储系统里面进行同步)
  12. . Failure Detection and Diagnostics(传感器错误的定义;传感器错误;传感数据错误;算法错误)
  13. How to Deal with Normal-Abnormal?(极端天气;紧急情况;工作区work zone?拥挤情况和延迟)
  14. Cyberattack Protection(通过多个传感设备协同纠正其中某个传感设备错误;量子计算发展导致现有技术难以保证数据安全;TEE受限的物理存储空间和执行性能)
  15. Vehicle Operating System(与嵌入式系统的兼容性;鲁棒性开源的操作系统)
  16. Energy Consumption
  17. Cost
  18. How to Benefit from Smart Infrastructure? (找到停车位;交通信息共享;任务卸载)
  19. Dealing with Human Drivers (人擅长感知环境;机器适合控制和信息处理,如何两者之间交互)
  20. Experimental Platform
  21. Physical Worlds Coupling Autonomous (需要结合复杂的交通情况和个人习惯,如个人习惯,急刹还是缓慢)

Prabhjot Kaur, Samira Taghavi, Zhaofeng Tian, Weisong Shi: A Survey on Simulators for Testing Self-Driving Cars. CoRR abs/2101.05337 (2021)

主要介绍了一些网联车的模拟平台,比较其优缺点
在这里插入图片描述

Sidi Lu, Xin Yuan, Aggelos K. Katsaggelos, Weisong Shi: Reinforcement Learning for Adaptive Video Compressive Sensing. CoRR abs/2105.08205 (2021)
在这里插入图片描述
通过强化学习的方法自适应压缩视频的快照

Jinghui Liao, Fengwei Zhang, Wenhai Sun, Weisong Shi: Speedster: A TEE-assisted State Channel System. CoRR abs/2104.01289 (2021)

针对区块链技术中状态信道网络的可扩展性问题、高昂的交易费用和低交易吞吐,提出了Speedster,一个基于帐户的状态通道系统,旨在解决上述问题。为此,Speedster利用最新开发的安全硬件创建无争议的认证渠道,可在区块链外高效运营。Speedster是完全分散的,提供更好的隐私保护。它支持快速的本机多方契约执行,这在以前启用TEE的信道网络中是不存在的。与Lightning网络相比,Speedster将吞吐量提高了约10000× 在网络规模相当的情况下,生成的链上数据减少了97%。

每个用户都创建一个由enclave(可信执行环境(TEE)的一个实例)保护的链外帐户并为其提供资金。随着Speedster将区块链的链上信任转移到enclaves的链下信任,我们显著降低了设计复杂性,以实现大量创新,例如多方渠道,以及用于渠道机密性、真实性、最终确定和争议解决的轻量级协议。Speedster在安全性、性能和功能方面优于传统的状态通道网络。

Quyuan Luo, Changle Li, Tom H. Luan, Weisong Shi: Collaborative Data Scheduling for Vehicular Edge Computing via Deep Reinforcement Learning. IEEE Internet Things J. 7(10): 9637-9650 (2020)

随着自动驾驶技术的发展,在可预见的未来,人们对数据通信以及联网和自动化车辆的计算卸载的需求将激增。在通信和计算能力有限的情况下,如何有效地调度网络中资源的使用,使其达到最佳利用率是一个基础性的研究课题。在本文中,我们通过联合考虑用于数据调度的通信和计算资源来解决这个问题。具体来说,我们研究了车辆边缘计算(VEC),其中边缘计算使路边单元(RSU)部署在道路沿线,为车辆提供数据带宽和计算卸载。此外,车辆之间可以通过数据中继和协作计算通过车辆对车辆(V2V)通信进行协作。在此基础上,提出了一个包含通信、计算、缓存和协同计算的统一框架,并提出了一种在保证应用延迟约束的前提下,使系统范围内的数据处理成本最小化的协同数据调度方案。为了得到数据调度的最优策略,我们进一步将数据调度建模为一个深度强化学习问题,该问题由一个单独目标Q网络的增强型深度Q网络(DQN)算法求解。通过大量的仿真,验证了该方案的有效性。

Jie Tang, Shaoshan Liu, Liangkai Liu, Bo Yu, Weisong Shi: LoPECS: A Low-Power Edge Computing System for Real-Time Autonomous Driving Services. IEEE Access 8: 30467-30479 (2020)

Challenge:
首先,这些计算密集型服务是由复杂的管道组成的,并且总是有严格的实时性要求。
其次,传感器数据形成一个时间序列,彼此独立。
最后,这样的车辆计算系统在电池上运行时能量预算非常有限。
Contribution:
首先,我们开发了一个异构资源感知的运行时层(Heterogeneity-Aware Runtime Layer),充分利用车辆的异构计算资源来满足自主驾驶应用的实时性要求;其次,我们开发了一个车辆边缘协调器来动态地将车辆任务卸载到edge cloudlet,以延长电池寿命的方式进一步优化用户体验;第三,我们成功地将这些组件集成到LoPECS系统中,并在nvidiajetsontx1上实现。

在这里插入图片描述
在这里插入图片描述

论文整理归纳分类:

网联车综述类

  1. Liangkai Liu, Sidi Lu, Ren Zhong, Baofu Wu, Yongtao Yao, Qingyang Zhang, Weisong Shi: Computing Systems for Autonomous Driving: State of the Art and Challenges. IEEE Internet Things J. 8(8): 6469-6486 (2021)
  2. Prabhjot Kaur, Samira Taghavi, Zhaofeng Tian, Weisong Shi: A Survey on Simulators for Testing Self-Driving Cars. CoRR abs/2101.05337 (2021)
  3. Liangkai Liu, Baofu Wu, Weisong Shi: A Comparison of Communication Mechanisms in Vehicular Edge Computing. HotEdge 2020

网联车设计类

  1. Qingyang Zhang, Hong Zhong, Jie Cui, Lingmei Ren, Weisong Shi: AC4AV: A Flexible and Dynamic Access Control Framework for Connected and Autonomous Vehicles. IEEE Internet Things J. 8(3): 1946-1958 (2021)
  2. Sidi Lu, Xin Yuan, Aggelos K. Katsaggelos, Weisong Shi: Reinforcement Learning for Adaptive Video Compressive Sensing. CoRR abs/2105.08205 (2021)
  3. Quyuan Luo, Changle Li, Tom H. Luan, Weisong Shi: Collaborative Data Scheduling for Vehicular Edge Computing via Deep Reinforcement Learning. IEEE Internet Things J. 7(10): 9637-9650 (2020)
  4. Jie Tang, Shaoshan Liu, Liangkai Liu, Bo Yu, Weisong Shi: LoPECS: A Low-Power Edge Computing System for Real-Time Autonomous Driving Services. IEEE Access 8: 30467-30479 (2020)
  5. Quyuan Luo, Changle Li, Tom H. Luan, Weisong Shi: EdgeVCD: Intelligent Algorithm-Inspired Content Distribution in Vehicular Edge Computing Network. IEEE Internet Things J. 7(6): 5562-5579 (2020)
  6. Sidi Lu, Xin Yuan, Weisong Shi: Edge Compression: An Integrated Framework for Compressive Imaging Processing on CAVs. SEC 2020: 125-138
  7. Zheng Dong, Yan Lu, Guangmo Tong, Yuanchao Shu, Shuai Wang, Weisong Shi: WatchDog: Real-time Vehicle Tracking on Geo-distributed Edge Nodes. CoRR abs/2002.04597 (2020)

物联网类

  1. Qingyang Zhang, Hong Zhong, Weisong Shi, Lu Liu: A trusted and collaborative framework for deep learning in IoT. Comput. Networks 193: 108055 (2021)
  2. Bodong Shang, Shiya Liu, Sidi Lu, Yang Yi, Weisong Shi, Lingjia Liu: A Cross-Layer Optimization Framework for Distributed Computing in IoT Networks. SEC 2020: 440-444
  3. Samira Taghavi, Weisong Shi: EdgeMask: An Edge-based Privacy Preserving Service for Video Data Sharing. SEC 2020: 382-387
  4. Huizi Xiao, Qingyang Zhang, Qingqi Pei, and Weisong Shi, Privacy-Preserving Neural Network Inference Framework via Homomorphic Encryption and SGX, in Proceedings of the the 41st IEEE International Conference on Distributed Computing Systems(ICDCS), July 7-10, 2021, Virtual.

模型压缩类

  1. Yanming Chen, Chao Li, Luqi Gong, Xiang Wen, Yiwen Zhang, Weisong Shi:A deep neural network compression algorithm based on knowledge transfer for edge devices. Comput. Commun. 163: 186-194 (2020)
  2. Francis McNamee, Schahram Dustdar, Peter Kilpatrick, Weisong Shi, Ivor T. A. Spence, Blesson Varghese: A Case For Adaptive Deep Neural Networks in Edge Computing. CoRR abs/2008.01814 (2020)
  3. Yanming Chen, Xiang Wen, Yiwen Zhang, and Weisong Shi, CCPrune: Collaborative Channel Pruning for Learning Compact Convolutional Networks, Neurocomputing, Vol. 451, September 2021, pp. 35-45.

能耗感知、资源分配类

  1. Congfeng Jiang, Tiantian Fan, Honghao Gao, Weisong Shi, Liangkai Liu, Christophe Cérin, Jian Wan: Energy aware edge computing: A survey. Comput. Commun. 151: 556-580 (2020)
  2. Kaile Xiao, Zhipeng Gao, Weisong Shi, Xuesong Qiu, Yang Yang, Lanlan Rui: EdgeABC: An architecture for task offloading and resource allocation in the Internet of Things. Future Gener. Comput. Syst. 107: 498-508 (2020)
  3. Kaile Xiao, Weisong Shi, Zhipeng Gao, Congcong Yao, Xuesong Qiu: DAER: A Resource Preallocation Algorithm of Edge Computing Server by Using Blockchain in Intelligent Driving. IEEE Internet Things J. 7(10): 9291-9302 (2020)
  4. Sa Wang, Yan-Hai Zhu, Shan-Pei Chen, Tian-Ze Wu, Wen-Jie Li, Xusheng Zhan, Haiyang Ding, Weisong Shi, Yungang Bao: A Case for Adaptive Resource Management in Alibaba Datacenter Using Neural Networks. J. Comput. Sci. Technol. 35(1): 209-220 (2020)
  5. Mohit Kumar, Xingzhou Zhang, Liangkai Liu, Yifan Wang, Weisong Shi: Energy-Efficient Machine Learning on the Edges. IPDPS Workshops 2020: 912-921
  6. Quyuan Luo, Changle Li, Tom H. Luan and Weisong Shi, Self-Learning based Computation Offloading for Internet of Vehicles: Model and Algorithm, accepted by IEEE Transactions on Wireless Communications (TWC), March 2021.
  7. Quyuan Luo, Changle Li, Tom H. Luan and Weisong Shi, Minimizing the Delay and Cost of Computation Offloading for Vehicular Edge Computing, accepted by IEEE Transactions on Services Computing, March 2021.

存储类

  1. Jie Tang, Shaoshan Liu, Jie Cao, Dawei Sun, Bolin Ding, Jean-Luc Gaudiot, Weisong Shi: π-Hub: Large-scale video learning, storage, and retrieval on heterogeneous hardware platforms. Future Gener. Comput. Syst. 102: 514-523 (2020)
  2. Ruijun Wang, Liangkai Liu, Weisong Shi: HydraSpace: Computational Data Storage for Autonomous Vehicles. CIC 2020: 70-77
  3. Sidi Lu, Bing Luo, Tirthak Patel, Yongtao Yao, Devesh Tiwari, Weisong Shi: Making Disk Failure Predictions SMARTer! FAST 2020: 151-167
  4. Lanyu Xu, Arun Iyengar, Weisong Shi: CHA: A Caching Framework for Home-based Voice Assistant Systems. SEC 2020: 293-306

区块链

  1. Jinghui Liao, Fengwei Zhang, Wenhai Sun, Weisong Shi: Speedster: A TEE-assisted State Channel System. CoRR abs/2104.01289 (2021)
  • 3
    点赞
  • 11
    收藏
    觉得还不错? 一键收藏
  • 3
    评论

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论 3
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值