An intelligent transportation system adopting optical wireless communications has many advantages such as low cost, accurate positioning and high scalability, whereas it faces many technical challenges. Firstly, due to the reflection/absorption/scattering/refraction effects on the optical signals from atmospheric mixed and random medium, the conventional indoor channel model for optical wireless communications could not be directly applied. Secondly, the state-of-the-art positioning methods based on signal characteristics, including received signal strength, time difference of arrival, angle of arrival and etc., possess high complexity and low robustness, which could not satisfy the requirements of high-accuracy, real-time positioning. Finally, due to non-ideal optical characteristics of devices, fast and relative movement of massive transceiver arrays (LEDs and image sensors) and the effects from atmospheric mixed and random medium, the inter-user interference mechanism is still unknown. Based on deep learning theory, computation vision theory, dynamic programming theory and game theory, this project investigates the theory and key technology of outdoor optical wireless communications and positioning for intelligent transportation systems. Specifically, it aims at solving three key scientific problems: 1) Build up time-varying geometry propagation channel model under complex climate and realize robust signal recovery; 2) Propose low-latency and high-precision vehicular positioning schemes based on computation vision theory; 3) Realize high-efficiency and cooperative vehicle-to-infrastructure/vehicle-to-vehicle communications adopting optical imaging technologies. The outputs from this project will provide the theoretical and technical foundations for seamless integration of optical wireless communications and intelligent transportation systems.
基于室外无线光通信的智能交通系统具有成本低、定位准、扩展性强等优点,但存在诸多技术挑战。首先,由于室外无线光信号受到大气混杂随机介质的反射/吸收/散射/折射作用,传统室内信道模型无法直接使用;第二,基于信号特征(接收信号强度、到达时差、到达方向角等)分析的传统定位算法存在复杂度高、鲁棒性差等问题,难以满足高精度、实时定位的需求;最后,受限于器件非理想光学特性、大规模收发阵列(LED光源和图像传感器)相对位置快速变化以及大气混杂随机介质影响,用户间干扰作用机理尚不明确。针对上述问题,本项目以深度学习、计算视觉、动态规划、博弈论等理论为主要分析工具,研究面向智能交通的室外无线光通信定位融合理论与关键技术,包括:①复杂气候条件下的时变几何传播模型和鲁棒信息恢复;②基于计算视觉理论的低时延、高精度车辆定位;③基于光学成像技术的协同高效通信,为无线光通信和智能交通系统有机结合夯实理论与技术基础。
基于室外无线光通信的智能交通系统具有成本低、定位准、扩展性强等优点,但存在诸多技术挑战。首先,由于室外无线光信号受到大气混杂随机介质的反射/吸收/散射/折射作用,传统室内信道模型无法直接使用;第二,基于信号特征(接收信号强度、到达时差、到达方向角等)分析的传统定位算法存在复杂度高、鲁棒性差等问题,难以满足高精度、实时定位的需求;最后,受限于器件非理想光学特性、大规模收发阵列(LED光源和图像传感器)相对位置快速变化以及大气混杂随机介质影响,用户间干扰作用机理尚不明确。针对上述问题,本项目以深度学习、计算视觉、动态规划、博弈论等理论为主要分析工具,研究面向智能交通的室外无线光通信定位融合理论与关键技术,包括:①复杂气候条件下的时变几何传播模型和鲁棒信息恢复;②基于计算视觉理论的低时延、高精度车辆定位;③基于光学成像技术的协同高效通信,为无线光通信和智能交通系统有机结合夯实理论与技术基础。项目执行期间,提出:①基于相干空间的无线光信道特征分析方法;②基于MIMO的室外无线光通信网络架构;③异构无线光网络的用户接入决策方法;④基于回归神经网络的无线光定位方法;⑤基于延时优化的车辆间通信链路管理方法;⑥车辆与路边基础设施间的跨时隙链路资源管理方法;⑦基于无定形小区的无线光通信干扰管理方法;⑧异构无线光车联网的分布式拓扑控制方法。在IEEE Transactions等期刊上发表SCI收录论文8篇;在IEEE Globecom等国际会议上发表EI收录论文4篇,申请中国发明专利5项(其中授权3项),项目负责人于2020年底当选IEEE Fellow、以第一完成人获2019年中国通信学会科学技术奖二等奖(自然科学类),圆满完成了任务书规定的研究目标。
{{i.achievement_title}}
数据更新时间:2023-05-31
4PAM-FTN大气光传输系统在弱湍流信道中的误码性能
无线充电电动汽车V2G模式下光储直流微电网能量管理策略
Gamma-Gamma湍流信道下广义空时脉冲位置调制
BDS/QZSS及其组合系统在中国和日本及周边地区的定位性能评估
Ordinal space projection learning via neighbor classes representation
融合能量收集的室外可见光通信理论与方法研究
面向智能交通的雷达通信融合方法与优化设计
无线传感器网络中面向弱测距的定位理论与关键技术
面向动态复杂无线环境基于多域无线信息融合的认知协作定位与跟踪模型