基于时空图注意力网络的车辆多模态轨迹预测模型
陈文强,王东丹,朱文英,汪勇杰,王涛

Vehicle multimodal trajectory prediction model based on spatio-temporal graph attention network
Wenqiang CHEN,Dongdan WANG,Wenying ZHU,Yongjie WANG,Tao WANG
表 1 不同模型在5 s预测范围内的均方根误差
Tab.1 RMSE for different models in 5-second forecast range
数据集RMSE
tf/sS-LSTM[16]CS-LSTM[15]S-GAN[14]PIP[18]STDAN[17]STGAMT(w/o IFE)STGAMT
HighD10.220.220.200.170.190.070.07
20.620.610.570.520.270.190.19
31.271.241.141.050.480.350.32
42.152.101.901.760.910.650.61
53.143.272.912.631.661.201.14
NGSIM10.650.610.570.550.420.220.21
21.311.271.321.181.010.800.78
32.162.092.221.941.691.521.49
43.253.103.262.882.562.472.40
54.554.374.404.043.673.693.58