Location Awareness Accuracy Improvement for Full-Mesh IoT Networks for CAV and 5G applications

位置感知设备是能够主动或被动地确定其相对于一个或多个已知参考的物理位置的设备。它们在互联和自动驾驶汽车（CAV）和5G通信网络领域发挥着至关重要的作用，因为下一代物联网（IoT）设备预计将具有可靠，高精度的位置检测能力。例如，向自动驾驶汽车的过渡将部分依靠所谓的“车辆到车辆”（V2 C）通信，其中车辆能够相互无线交换有关其周围环境的信息-它们的位置，速度等。位置信息在手持和可穿戴物联网设备中的重要性已得到充分确认，最近，作为国家应对COVID-19大流行的一部分，接触追踪应用程序的重要性进一步加强。虽然近年来取得了长足的进步，但在位置感知技术的可靠性能方面仍然存在相当大的挑战，必须克服这些挑战，才能在复杂和安全关键的现实环境中成功部署，特别是当设备高速行驶时。该研究项目旨在研究和开发敏捷人工智能（AI）和机器学习（ML）算法，以解决相互传感器干扰和位置感知准确性等问题，特别是涉及自动驾驶和远程监控车辆的可靠性。我们可以将该项目概括为三个主要目标。首先，我们将为全网状位置感知系统开发AI/ML算法，该系统考虑了天线特性、传播信道和复杂现实环境中的干扰等因素。我们将继续将这些算法嵌入到定制的位置感知物联网测试平台中，能够对一系列环境进行建模，并开发用于演示和测试目的的图形工具。我们建议使用一个网络的商业可用的“群”无线电模块-一个网络的独立节点能够执行飞行时间测距到所有连接的节点，以确定使用宽带UWB和CHIRP通信协议的实时位置。最后，我们将开发评估和验证方法，以提高移动对象的位置感知精度。关于EPSRC的研究重点，该项目与研究理事会的重点“普及和无处不在的计算”密切相关;开发环境和位置感知是将物联网智能技术集成到交通和城市基础设施等领域的重要组成部分，并补充EPSRC的C2和C3连接成果，以实现物联网的转型发展，并提供智能和在交通运输领域的互联解决方案。创新人工智能算法的开发补充了对人工智能技术以及“ICT网络和分布式系统”的关注，通过我们使用全网状5G网络来促进位置估计，以及我们对相互传感器干扰等缓解因素的研究。该项目与英国国家测量标准实验室国家物理实验室（NPL）合作，研究支持政府关于CAV系统的主题，重点是成功引入可靠的无人驾驶和驾驶辅助车辆。人工智能，传感器和控制系统正在彻底改变陆地，海洋和航空运输服务，自动化是政府的优先事项，有可能破坏车辆所有权模式。这与向电动汽车的过渡相结合，是在交通运输领域实现净零二氧化碳目标的重要一步，我们相信我们的工作具有强大的潜力，可以增加该领域基于自动化技术的整合。

Location-aware devices are devices capable of determining their physical location, actively or passively, in relation to a known reference or references. They play a vital role within the areas of connected and autonomous vehicles (CAV) and 5G communication networks, as next-generation Internet of Things (IoT) devices are expected to have reliable, high-accuracy location detection capabilities. For example, the transition to autonomous vehicles will be underpinned in part by so-called "Vehicle-to-Vehicle" (V2C) communication, in which vehicles are able to wirelessly exchange information about their surroundings - their location, speed, etc. - with each other. The importance of location information in handheld and wearable IoT devices is well established and was recently reinforced by the importance of contact tracing apps as part of the national response to the COVID-19 pandemic. While recent years have made strong advancements, there are still considerable challenges in the reliable performance of location-aware technology that must be overcome to enable successful deployment in complex and safety-critical real-world environments, particularly when devices are travelling at high speeds. This research project aims to perform research and development regarding agile artificial intelligence (AI) and machine learning (ML) algorithms for addressing issues such as mutual sensor interference and location awareness accuracy, particularly as it pertains to the reliability of self-driving and remotely monitored vehicles. We can summarise the project in terms of three main objectives. Firstly, we will develop AI/ML algorithms for a full-mesh location-aware system, which considers factors such as antenna characteristics, propagation channels, and interference within complex real-world environments. We will proceed to embed these algorithms into bespoke location-aware IoT testbeds, capable of modelling a range of environments, and develop graphical tools for demonstration and testing purposes. We propose to use a network of commercially available 'Swarm' radio modules - a network of independent nodes capable of performing time-of-flight ranging to all connected nodes to determine real-time location using broadband UWB and CHIRP communication protocols. Finally, we will develop evaluation and validation methods for location awareness accuracy improvement for moving objects. With respect to the EPSRC's research priorities, this project aligns closely with the research council's focus on 'Pervasive and ubiquitous computing'; developing context and location awareness is an important part of integrating IoT smart technologies into areas such as transport and urban infrastructure, and complements EPSRC's C2 and C3 Connected Outcomes to achieve transformational development in the Internet of Things and deliver intelligent and connected solutions in the transport sector. The development of innovative AI algorithms complements the focus on Artificial Intelligence technologies, as well as 'ICT networks and distributed systems', through our use of full-mesh 5G networks to facilitate location estimation, and our research into mitigating factors such as mutual sensor interference. The project is in collaboration with the National Physical Laboratory (NPL), the UK's national measurement standards laboratory, and the research supports a government theme on CAV systems, focusing on the successful introduction of reliable driverless and driver-assisted vehicles. AI, sensor and control systems are revolutionising terrestrial, marine and aviation transport services, and automation is a government priority with the potential to disrupt models of vehicle ownership. This, combined with the transition to electric vehicles, is an important step towards achieving net-zero CO2 targets within the transport sector, and we believe our work has strong potential to increase the integration of automation-based technology in this area.