智能网联车辆混行交通流中灯语意图识别模型研究

为使混行交通流下智能网联车辆(Connected and Automated Vehicles, CAV)实现对人工驾驶车辆(Human-driven Vehicle, HV)前照灯灯语意图(Vehicle Headlights Intention, VHI) 的识别，弥补车对车(Vehicle to Vehicle, V2V)和鸣笛意图识别技术的不足，更好地与HV交互沟通，提出CAV对HV的VHI识别模型.模型包括：灯光感知、光数据处理、VHI识别3个模块，灯光感知模块通过RGB(Red-Green-Blue, RGB)和HSV(Hue-Saturation-Value, HSV)颜色空间感知前照灯(Vehicle Headlights, VH)，采用KLT(Kanade-Lucas-Tomasi Tracking，KLT)和车辆匹配算法定位跟踪发出灯语的HV；光数据处理模块采用光通道增益算法计算光辐射通量变化； VHI识别模块基于双层隐马尔可夫模型(Double-layer Hidden Markov Model，DHMM)辨识VH 闪烁次数和HV行驶状态，实现VHI识别.在3种灯语示意典型场景下的实验结果表明：1 s内 VH感知准确率为96.8%，定位跟踪精度小于1°，VHI识别率为96.6%，满足混行交通环境下 CAV对HV驾驶意图的识别要求，基本保证实时性，为混行交通流中CAV自动驾驶决策提供理论依据.

Vehicle Headlights Intention Recognition Model for Connected and Automated Vehicles in Mixed Traffic Flow

This study proposes a Vehicle Headlights Intention (VHI) recognition model to improve the communication between the Connected and Automated Vehicles (CAV) and Human- driven Vehicles (HV). The VHI recognition model is consist of light perception module, optical data processing module, and VHI recognition module. The light perception module is able to locate and track the HV that sent a light signal through the RedGreen-Blue (RGB) and Hue-Saturation-Value (HSV) color space, the Kanade-Lucas-Tomasi Tracking (KLT), and the vehicle matching algorithm. The optical data processing module calculates the optical radiant flux using optical channel gain algorithm. The VHI recognition module identifies the number of headlight flashing and vehicle driving status based on the Double- layer Hidden Markov Model(DHMM). The experimental results from three typical VHI scenarios indicate that the average accuracy of vehicle headlights perception is 96.8%, and the error of positioning and tracking is within 1 degree. The 1-second VHI recognition rate reaches 96.6%, which enables the driving intention recognition of CAVs and provides the basis for the automated driving decision of the CAV in mixed traffic flow.