融合图卷积网络与社交池化的多模态轨迹预测模型
|
|
赵庆慧,崔鑫,张艺炜,陈燕
|
Multimodal trajectory prediction model integrating graph convolutional networks and social pooling
|
|
Qinghui ZHAO,Xin CUI,Yiwei ZHANG,Yan CHEN
|
|
| 表 2 不同轨迹预测模型在HighD数据集上的性能对比 |
| Tab.2 Performance comparison of different trajectory prediction models on dataset HighD |
|
| 模型 | RMSE/m | | NLL | | tp=1 s | tp=2 s | tp=3 s | tp=4 s | tp=5 s | 平均值 | | tp=1 s | tp=2 s | tp=3 s | tp=4 s | tp=5 s | 平均值 | | CV | 0.33 | 0.78 | 1.62 | 2.43 | 3.67 | 1.766 | | 1.94 | 3.09 | 4.85 | 6.12 | 7.03 | 4.606 | | PiP | 0.24 | 0.68 | 1.34 | 2.19 | 3.42 | 1.574 | | 0.45 | 2.27 | 3.34 | 4.20 | 4.76 | 3.004 | | WSiP | 0.22 | 0.59 | 1.21 | 2.05 | 3.04 | 1.422 | | 0.34 | 1.83 | 2.72 | 3.49 | 4.25 | 2.526 | | HAN | 0.18 | 0.44 | 1.05 | 1.72 | 2.87 | 1.252 | | 0.27 | 1.24 | 2.37 | 3.06 | 3.95 | 2.178 | | HLTP++ | 0.19 | 0.42 | 1.11 | 1.64 | 2.75 | 1.222 | | 0.29 | 1.38 | 2.59 | 3.04 | 3.87 | 2.234 | | DRBP | 0.18 | 0.49 | 1.11 | 1.62 | 2.72 | 1.224 | | 0.24 | 1.27 | 2.44 | 2.93 | 3.69 | 2.114 | | GC-LSTM | 0.15 | 0.37 | 0.99 | 1.54 | 2.65 | 1.140 | | 0.23 | 1.07 | 2.04 | 2.74 | 3.53 | 1.922 |
|
|
|