With the burgeoning of Sensing, Edge Computing, AI, and 5G technologies, we envision future vehicles will serve as a computing platform for a variety of services like advanced driver assistance system (ADAS) and Autonomous Driving (AD), remote real-time diagnostics, in-vehicle infotainment, and a variety of third-party services, such as public safety and weather alert. As the amount of data generated by these services on one vehicle increases dramatically and has the potential to reach four terabytes per day, it is a huge challenge to guarantee performance through the limited on-board computing power. Therefore, an efficient data analytics platform, including both hardware and software, is needed to enable future vehicle computing. In partnership with our industry collaborators,the CAR lab at Wayne State University aims to design and implement enabling technologies, including edge computing/communication systems, data analytics and applications, secure trusted execution environment, privacy preserving models and tools, to realize the vision of connected and autonomous driving.
Given the unique position of Detroit as the hub of the North American automobile industry, we have founded
annual workshop in this region.
Many automobile companies are working on their proprietary computing platform, however, an open platform that provide interfaces for researchers and developers is missing. In this project, we aim to develop such a platform that provides a full stack solution and dynamically senses services status and adjusts services to guarantee service quality and user experience. We are developing on an open vehicular data analytics platform (OpenVDAP),which consists of four components: heterogeneous vehicle computing unit (VCU) , operation system (EdgeOS) driving data collector and integrator (DDI), and application library (libvdap).
heterogeneous vehicle computing unit:We intend to manage and utilize the heterogeneous computing hardware resources in a future vehicle, e.g., CPU, GPU and FPGA, as well as computing devices from passengers, to support real-time data processing. We also aim to manage the multiple communication components, e.g., 3G/4G/5G and DSRC, available in the vehicle.
edge operating system for vehicle: The core of OpenVDAP is an edge operating system for vehicles, call EdgeOSv, in which each service offers multiple execution pipelines in response to various network and computational constraints. EdgeOSv employs an elastic management model that automatically chooses an optimal pipeline to reduce the lowest end-to-end latency.
driving data collector and integrator: we build a driving data collector and integrator. This service will collect all data generated by the vehicle, including the data read by OBD reader and other sensors' data, such as dash camera. Moreover, this service will integrate related data from Internet, such as weather, road condition and so on.
application library: We provide a library for developers who want to build third-party services on our EdgeOSv. Using libvdap, developers can access all vehicle data by communicating with DDI service.
- Dalong Li, FCA
- Shaoshan Liu,PerceptIn
- Mu Qiao, IBM Research
- Haris Volos, DENSO International America, Inc.
- Zhifeng Yu, EdgeMind LLC.
- Liangkai Liu, Xingzhou Zhang, Mu Qiao, and Weisong Shi, SafeShareRide: Edge-based Attack Detection in Ridesharing Services, Technical Reprot MIST-TR-2018-004, March, 2018.
- Qingyang Zhang, Yifan Wang, Xingzhou Zhang,Liangkai Liu, Xiaopei Wu, Weisong Shi and Hong Zhong, OpenVDAP: An Open Vehicular Data Analytics Platform for CAVs, in Proceedings of the 38th IEEE International Conference on Distributed Computing Systems (ICDCS), Vision/Blue Sky Track, July 2-5, 2018, Vienna, Austria.
- Quan Zhang, Qingyang Zhang, Weisong Shi and Hong Zhong, Firework: Data Processing and Sharing for Hybrid Cloud-Edge Analytics, accepted by IEEE Transactions on Parallel and Distributed Systems (TPDS), February 2018.
- Qingyang Zhang, Quan Zhang, Weisong Shi and Hong Zhong, Distributed Collaborative Execution on the Edges and Its Application on AMBER Alert,accepted by IEEE Internet of Things Journal. Feb. 2018
- Kewei Sha, Wei Wei, Andrew T. Yang, Zhiwei Wang and Weisong Shi, On Security Challenges and Open Issues in Internet of Things, accepted by Future Generation Computer Systems, January 2018.
- Jie Tang, Shaoshan Liu, Jie Cao, Dawei Sun, Bolin Ding, Zhe Zhang, Jean-Luc Gaudiot and Weisong Shi, pi-Hub:Large-Scale Video Learning, Storage,and Retrieval on Heterogeneous Hardware Platforms, Technical Reprot MIST-TR-2017-012, November 2017.
- Shanhe Yi, Zijiang Hao, Qingyang Zhang, Quan Zhang, Weisong Shi and Qun Li, LAVEA: Latency-aware Video Analytics on Edge Computing Platform , in Proceedings of 2nd ACM/IEEE Symposium on Edge Computing (SEC), San Jose, Oct 12-14, 2017.
- Qingyang Zhang, Quan Zhang, Weisong Shi and Hong Zhong, Poster: Enhancing AMBER Alert using Collaborative Edges , in Proceedings of 2nd ACM/IEEE Symposium on Edge Computing ( SEC),San Jose, Oct 12-14, 2017.
- Zhenyu Ning, Fengwei Zhang, Weisong Shi and Larry Shi, Position Paper: Challenges Toward Securing Hardware-assisted Execution Environments, in Proceedings of Hardware and Architectural Support for Security and Privacy (HASP), in conjunction with ISCA, Toronto, Canada, June 25, 2017.
- Quan Zhang, Qingyang Zhang, Weisong Shi and Hong Zhong, Firework: Data Processing and Sharing for Hybrid Cloud-Edge Analytics, Technical Report MIST-TR-2017-002, January 2017.
- Qingyang Zhang, Zhifeng Yu, Weisong Shi and Hong Zhong, Demo Abstract: EVAPS: Edge Video Analysis for Public Safety, in Proceedings of 1st IEEE/ACM Symposium on Edge Computing (SEC), Washington DC, Oct 27-28, 2016.
- Quan Zhang, Xiaohong Zhang, Qingyang Zhang, Weisong Shi and Hong Zhong, Firework: Big Data Sharing and Processing in Collaborative Edge Environment, in Proceedings of 4th IEEE Workshop on Hot Topics in Web Systems and Technologies (HotWeb), Washington DC, Oct. 24-25,2016.
- Weisong Shi, Jie Cao, Quan Zhang, Youhuizi Li and Lanyu Xu, Edge Computing: Vision and Challenges, IEEE Internet of Things Journal, Vol. 3, No. 5, October 2016, pp. 637-646.