Stepwise Goal-Driven Networks for Trajectory Prediction

25 Mar 2021  ยท  Chuhua Wang, Yuchen Wang, Mingze Xu, David J. Crandall ยท

We propose to predict the future trajectories of observed agents (e.g., pedestrians or vehicles) by estimating and using their goals at multiple time scales. We argue that the goal of a moving agent may change over time, and modeling goals continuously provides more accurate and detailed information for future trajectory estimation. To this end, we present a recurrent network for trajectory prediction, called Stepwise Goal-Driven Network (SGNet). Unlike prior work that models only a single, long-term goal, SGNet estimates and uses goals at multiple temporal scales. In particular, it incorporates an encoder that captures historical information, a stepwise goal estimator that predicts successive goals into the future, and a decoder that predicts future trajectory. We evaluate our model on three first-person traffic datasets (HEV-I, JAAD, and PIE) as well as on three bird's eye view datasets (NuScenes, ETH, and UCY), and show that our model achieves state-of-the-art results on all datasets. Code has been made available at: https://github.com/ChuhuaW/SGNet.pytorch.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Trajectory Prediction ETH/UCY SGNet ADE-8/12 0.18 # 2
FDE-8/12 0.35 # 7
Trajectory Prediction HEV-I SGNet ADE(0.5) 6.28 # 1
ADE(1.0) 11.35 # 1
ADE(1.5) 18.27 # 1
FDE(1.5) 39.86 # 1
FIOU(1.5) 0.63 # 1
Multi-future Trajectory Prediction JAAD SGNet MSE(0.5) 37 # 1
MSE(1.0) 86 # 1
MSE(1.5) 197 # 1
C_MSE(1.5) 146 # 1
CF_MSE(1.5) 443 # 1
Trajectory Prediction JAAD SGNet MSE(0.5) 82 # 1
MSE(1.0) 328 # 1
MSE(1.5) 1049 # 1
C_MSE(1.5) 996 # 1
CF_MSE(1.5) 4076 # 1
Multi-future Trajectory Prediction PIE SGNet MSE(0.5) 16 # 1
MSE(1.0) 39 # 1
MSE(1.5) 88 # 1
C_MSE(1.5) 66 # 1
CF_MSE(1.5) 206 # 1
Trajectory Prediction PIE SGNet MSE(0.5) 34 # 1
MSE(1.0) 133 # 1
MSE(1.5) 442 # 1
C_MSE(1.5) 413 # 1
CF_MSE(1.5) 1761 # 1

Methods


No methods listed for this paper. Add relevant methods here