TASCC: Secure Cloud-based Distributed Control (SCDC) Systems for Connected Autonomous Cars

TASCC：用于联网自动驾驶汽车的安全的基于云的分布式控制 (SCDC) 系统

• 批准号: EP/N01300X/1
• 负责人: Mehrdad Dianati
• 金额: $ 331.43万
• 依托单位: University of Surrey
• 依托单位国家: 英国
• 项目类别: Research Grant
• 财政年份: 2016
• 资助国家: 英国
• 起止时间: 2016 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=EP%2FN01300X%2F1
• 关键词: TASCC; Secure; Cloud; based; Distributed

Automotive industry and the consumers are eager for smart features on new cars and more efficient vehicles. Modern cars are not considered as mere means for travelling from point A to B anymore, but rather smart systems that offer personalised services and have the capability to adapt to the user's preferences and needs. They are expected to become intelligent agents that learn from their environments and exploit various sources of information to become increasingly autonomous systems that relieve the driver from tedious tasks, such as parking, and improve safety, efficiency, and desirability of the future cars. From a wider angle, today's land transportation systems claim about 1.3 million lives and 7 million injuries in road accidents, according to a recent report by CISCO. The increasing number of cars results in traffic jams costing about 90 billion of lost hours for the drivers and the passengers. In addition, transportation accounts for about 26% of the total greenhouse gas emission from human activities. While public transport can help, cars remain to be the desired means of transport according to a recent report by the Department of Transport in 2014. These market forces in addition to the environmental, economic and social impacts of transport systems demand a timely and transformative research to rethink the automotive control systems and revolutionise vehicle design for future cars. There have been two trends towards this objective in the past decade: in the one hand the research in autonomous systems, inspired by unmanned space vehicles, gave birth to driver-less concept cars such as Google robotic car; on the other hand, modern wireless communications enabled cars to talk to each other and the roadside infrastructures, resulting in the concept of connected cars. However, driver-less cars remain to be too expensive for commercial vehicles (Google's cars cost about £100,000 only for sensing equipment) and connected vehicles can offer little if not properly integrated into smart and autonomous features. This ambitious research is defined by a number of world-class academic institutions and leading industrial partners to work with Jaguar Land Rover, a market leader in high end cars, to design and validate a framework that combines the power of connected vehicles concept with the notion of autonomous systems and build a novel platform for cost-effective deployment of autonomous features and ultimately realisation of connected and fully autonomous cars. This can be made possible thanks to modern wireless technologies and the power of cloud computing that allows sharing expensive computing resources (hence, reducing costs per vehicle) and provides access to information that are only available on the cloud. To realise the ambition of the project, a number of key challenges in the areas of ultra-low-latency wireless technologies, cloud computing, distributed control systems, and human interaction issues will be addressed in this project. In addition, potential security threats will be identified and analysed to assess the potential risks for the public and reputational damage for car manufacturers should such technologies be commercialised. At the end of the project, the technical solutions will be integrated into a single framework and will be validated by example applications, characterising technical and service-level performance of the framework, and providing a basis for the future direction of enhanced automated services. While the objective here is to ultimately enable affordable driver-less cars, in the short term, this project aims to enable a number of demonstrable autonomous features in a test environment.

汽车工业和消费者都渴望在新的汽车和更高效的车辆上具有智能功能。现代汽车不再仅仅被认为是从A点到B点的交通工具，而是提供个性化服务并能够适应用户偏好和需求的智能系统。它们有望成为智能代理，从环境中学习，并利用各种信息源成为越来越自主的系统，将驾驶员从繁琐的任务中解放出来，例如停车，并提高未来汽车的安全性，效率和可取性。根据思科最近的一份报告，从更广泛的角度来看，今天的陆地交通系统在道路事故中造成约130万人死亡，700万人受伤。汽车数量的增加导致交通堵塞，司机和乘客损失了大约900亿小时。此外，交通运输约占人类活动温室气体排放总量的26%。虽然公共交通可以提供帮助，但根据交通部2014年的一份最新报告，汽车仍然是理想的交通工具。除了运输系统的环境、经济和社会影响之外，这些市场力量还需要进行及时的变革性研究，以重新思考汽车控制系统并彻底改变未来汽车的车辆设计。在过去十年中，朝着这一目标有两个趋势：一方面，受无人驾驶航天器的启发，自主系统的研究催生了无人驾驶概念汽车，如谷歌机器人汽车;另一方面，现代无线通信使汽车能够相互交谈并与路边基础设施交谈，从而产生了联网汽车的概念。然而，无人驾驶汽车对于商用车来说仍然过于昂贵（谷歌的汽车仅传感设备就花费了约10万英镑），而联网汽车如果没有适当地集成到智能和自动驾驶功能中，则几乎无法提供任何功能。这项雄心勃勃的研究是由一些世界级的学术机构和领先的工业合作伙伴定义的，与高端汽车市场领导者捷豹路虎合作，设计和验证一个框架，将互联车辆概念的力量与自动驾驶系统的概念相结合，并建立一个新的平台，以经济有效地部署自动驾驶功能，并最终实现互联和完全自动驾驶的汽车。由于现代无线技术和云计算的强大功能，这一点成为可能，云计算允许共享昂贵的计算资源（因此，降低了每辆车的成本），并提供了对仅在云上可用的信息的访问。为了实现该项目的宏伟目标，该项目将解决超低延迟无线技术、云计算、分布式控制系统和人机交互问题等领域的一些关键挑战。此外，还将识别和分析潜在的安全威胁，以评估这些技术商业化后对公众的潜在风险和对汽车制造商的声誉损害。在项目结束时，这些技术解决方案将被纳入一个单一的框架，并将通过示例应用程序进行验证，说明框架的技术和服务水平性能，并为今后加强自动化服务的方向奠定基础。虽然目标是最终实现负担得起的无人驾驶汽车，但在短期内，该项目旨在在测试环境中实现一些可演示的自动驾驶功能。

项目成果

Emerging Privacy Challenges and Approaches in CAV Systems

• DOI: 10.1049/cp.2019.0141
• 发表时间: 2019
• 期刊: Living in the Internet of Things (IoT 2019)
• 作者: U. Atmaca;Carsten Maple;M. Dianati

Optimising driver profiling through behaviour modelling of in-car sensor and global positioning system data

通过车内传感器和全球定位系统数据的行为建模优化驾驶员分析

• DOI: 10.1016/j.compeleceng.2021.107047
• 发表时间: 2021
• 期刊: Computers & Electrical Engineering
• 影响因子: 4.3
• 作者: Ahmadi-Assalemi G

A Taxonomy and Survey of Edge Cloud Computing for Intelligent Transportation Systems and Connected Vehicles

• DOI: 10.1109/tits.2021.3084396
• 发表时间: 2021-06-04
• 期刊: IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
• 影响因子: 8.5
• 作者: Arthurs, Peter;Gillam, Lee;Mouzakitis, Alexandros
• 通讯作者: Mouzakitis, Alexandros

Cooperative Object Classification for Driving Applications

• DOI: 10.1109/ivs.2019.8813811
• 发表时间: 2019-06
• 期刊: 2019 IEEE Intelligent Vehicles Symposium (IV)
• 作者: Eduardo Arnold;Omar Y. Al-Jarrah;M. Dianati;Saber Fallah;David Oxtoby;A. Mouzakitis

Fast and Robust Registration of Partially Overlapping Point Clouds

• DOI: 10.1109/lra.2021.3137888
• 发表时间: 2021-12
• 期刊: IEEE Robotics and Automation Letters
• 影响因子: 5.2
• 作者: Eduardo Arnold;Sajjad Mozaffari;M. Dianati