/Res-GCNN

A Lightweight Residual Graph CNN for Pedestrians Trajectory Prediction

Res-GCNN

We present a novel network called Residual Graph Convolutional Neural Network (Res-GCNN) for pedestrians trajectory prediction. We design Res-GCNN for onboard application in autonomous vehicles, therefore, real-time performance is a key factor.
The Res-GCNN models the interactive behaviors of pedestrians by using the adjacent matrix of the constructed graph for the current scene. Though the proposed Res-GCNN is quite lightweight with only about 6.4 kilo parameters which outperforms all other methods in terms of parameters size, our experimental results show an improvement over the state of art by 13.3% on the Final Displacement Error (FDE) which reaches 0.65 meter. As for the Average Displacement Error (ADE), we achieve a suboptimal result (the value is 0.37 meter), which is also very competitive. The Res-GCNN is evaluated in the platform with an NVIDIA GeForce RTX1080Ti GPU, and its mean inference time of the whole dataset is only about 2.2 microseconds. Compared with other methods, the proposed method shows strong potential for onboard application accounting for forecasting accuracy and time efficiency.
Detailed presentation is available at http://arxiv.org/abs/2011.09214

Model Architecture

The proposed Res-GCNN contains two components, FECNN and TPCNN. FECNN plays the role of extracting spatio-temporal inter-action feature, while TPCNN is designed to forecast the future trajectories. The Res-GCNN is composed of several FECNN layers and several TPCNN layers.

image

The ADE / FDE metrics of Res-GCNN compared with other methods

The positions are observed every 0.4 second. The same as our superior counterparts, we use 8 historical data frames to forecast the coming 12 data frames.

Model\Dataset ETH HOTEL UNIV ZARA1 ZARA2 AVG
Linear 1.33 / 2.94 0.39 / 0.72 0.82 / 1.59 0.62 / 1.21 0.77 / 1.48 0.79 / 1.59
SR-LSTM-2 0.63 / 1.25 0.37 / 0.74 0.51 / 1.10 0.41 / 0.90 0.32 / 0.70 0.45 / 0.94
S-LSTM 1.09 / 2.35 0.79 / 1.76 0.67 / 1.40 0.47 / 1.00 0.56 / 1.17 0.72 / 1.54
S-GAN-P 0.87 / 1.62 0.67 / 1.37 0.76 / 1.52 0.35 / 0.68 0.42 / 0.84 0.61 / 1.21
SoPhie 0.70 / 1.43 0.76 / 1.67 0.54 / 1.24 0.30 / 0.63 0.38 / 0.78 0.54 / 1.15
CGNS 0.62 / 1.40 0.70 / 0.93 0.48 / 1.22 0.32 / 0.59 0.35 / 0.71 0.49 / 0.97
PIF 0.73 / 1.65 0.30 / 0.59 0.60 / 1.27 0.38 / 0.81 0.31 / 0.68 0.46 / 1.00
STSGN 0.75 / 1.63 0.63 / 1.01 0.48 / 1.08 0.30 / 0.65 0.26 / 0.57 0.48 / 0.99
GAT 0.68 / 1.29 0.68 / 1.40 0.57 / 1.29 0.29 / 0.60 0.37 / 0.75 0.52 / 1.07
Social-BiGAT 0.69 / 1.29 0.49 / 1.01 0.55 / 1.32 0.30 / 0.62 0.36 / 0.75 0.48 / 1.00
Social-STGCNN 0.64 / 1.11 0.49 / 0.85 0.44 / 0.79 0.34 / 0.53 0.30 / 0.48 0.44 / 0.75
Meituan 0.39 / 0.79 0.51 / 1.05 0.25 / 0.56 0.30 / 0.61 0.36 / 0.73 0.36 / 0.75
Res-GCNN 0.65 / 1.12 0.31 / 0.50 0.38 / 0.72 0.28 / 0.50 0.24 / 0.41 0.37 / 0.65

Results of Typical Patterns

Parallel walking

images

images

images

Collision avoidance

images

images

images

Failure scenes

images

images

images

Complex scenarios

images

images

images