NeTS: Small: Infrastructure-free Robust Relative Localization of Vehicles on the Road

汽车行业正在经历一场巨大的变革，因为车辆正在准备控制远离人类的驾驶。这将减少事故发生的机会，无压力驾驶，增加乘客舒适性，提高燃油效率，减少旅行时间。实现这一自动驾驶汽车愿景的最关键信息之一是准确确定附近车辆的位置。该项目的总体目标是开发可靠的方法，这些方法实际上可以实现附近车辆的细粒度相对本地化，而不需要基础设施的支持。特别是，将利用车载传感能力和车辆之间的射频通信。该项目的主要特点如下：1)综合解决细粒度车辆定位问题，同时考虑到传统车辆是推出自动驾驶车辆的最关键障碍；2)创新的物理层技术，以实现高精度的相对定位；以及3)使用由各种传感器收集的噪声数据并仍然稳健地得出车辆位置的技术。该项目具有潜在的变革性，因为它解决了车辆准确相对定位的关键问题，同时考虑了实际挑战，这对自动驾驶技术至关重要。更广泛的影响包括加强本科生和研究生的课程。此外，在整个项目中，我们将有效地利用与汽车行业的现有合作，以获得关于目标的指导，并了解我们解决方案的可行性。1)基于频率对的单天线车辆分析：本文的创新之处在于提出了一种基于频率对分析的方法，可以很好地解决实际驾驶环境中的距离估计误差和变幻莫测的问题。2)在多天线车辆中利用空间分集：这一推力的关键创新在于它利用了多天线提供的空间分集来增加稳健性。3)传统车辆存在下的协同车辆地图构建：关键创新是将通过安装在参与车辆上的摄像头获得的噪声信息进行强有力的融合。

The automotive industry is undergoing a giant transformation as vehicles are gearing up to take control of driving away from humans. This will lead to reduced chance of accidents, stress-free driving, increased passenger comfort, increased fuel-efficiency and reduced travel time. One of the most critical pieces of information for enabling this vision of autonomous vehicles is accurate determination of locations of nearby vehicles. The overarching goal of this project is to develop robust approaches that are practically implementable for fine-grained relative localization of nearby vehicles with no support from the infrastructure. In particular, on-board sensing capabilities and RF (radio-frequency) communication between vehicles will be leveraged. The key distinguishing features of the project are as follows: 1) a comprehensive approach to the problem of fine-grained vehicular localization while considering legacy vehicles which is the most critical hurdle in rolling out autonomous vehicles; 2) innovative physical layer techniques to achieve high accuracy relative localization; and, 3) techniques to use noisy data collected by various sensors and still robustly derive the locations of the vehicles. The project is potentially transformative as it addresses the key question of accurate relative localization of vehicles while considering practical challenges, which is crucial for the autonomous driving technology. The broader impacts include enhancing undergraduate and graduate curriculum. In addition, existing collaboration with the automotive industry will be effectively leveraged for obtaining guidance on the objectives and understanding the feasibility of our solutions throughout the project.The specific inter-related research thrusts are as follows. 1) Frequency-Pair based Analysis for Single Antenna Vehicles: The key innovation is an approach that can robustly tackle the distance estimation error and the vagaries of a real driving environment using analysis of frequency-pairs. 2) Exploiting Spatial Diversity in Multi-Antenna Vehicles: The key innovation of this thrust is that it uses the spatial diversity offered by multiple antennas to increase robustness. 3) Collaborative Vehicle-Map Construction in Presence of Legacy Vehicles: The key innovation is to robustly fuse noisy information obtained through cameras mounted on the participating vehicles.