融合多车风格感知与交互特征的换道行为预测

韩泽一,王文璇,王元庆

Prediction of lane-changing behavior integrating multi-vehicle style perception and interaction features

Zeyi HAN,Wenxuan WANG,Yuanqing WANG

表 6 考虑驾驶风格的LSTM和2-CNN模型预测结果

Tab.6 Prediction results of LSTM and 2-CNN models considering driving style

T/s Acc/% P/% R/% F1/% LSTM 2-CNN LSTM 2-CNN LSTM 2-CNN LSTM 2-CNN 1.0 99.03 97.72 95.95 96.33 97.83 89.27 96.88 92.68 1.5 99.36 98.35 98.75 97.28 97.59 96.42 98.17 96.85 2.0 99.74 98.27 98.96 97.14 99.57 95.84 99.26 96.48 2.5 99.73 98.95 98.90 96.46 99.57 96.50 99.23 96.48 3.0 99.54 98.03 98.58 94.37 98.83 95.11 98.71 94.74 3.5 99.56 97.59 99.78 92.77 97.34 91.11 98.54 91.93 4.0 99.52 97.47 98.11 93.73 97.50 87.31 97.80 90.40 4.5 99.37 97.50 99.04 95.25 95.26 82.60 97.11 88.34 5.0 98.59 98.11 92.25 94.35 91.23 87.24 91.72 90.64