VisionNet: A Drivable-space-based Interactive Motion Prediction Network for Autonomous Driving

TL;DR

VisionNet is a CNN-based network that predicts future drivable spaces for autonomous vehicles by modeling interactions more effectively, leading to improved trajectory prediction accuracy in complex traffic scenarios.

Contribution

The paper introduces VisionNet, a novel approach transforming interaction modeling into drivable space estimation with an interactive loss for better supervision.

Findings

Outperforms existing models on public datasets

Effectively models complex interactions in traffic scenarios

Improves trajectory prediction accuracy

Abstract

The comprehension of environmental traffic situation largely ensures the driving safety of autonomous vehicles. Recently, the mission has been investigated by plenty of researches, while it is hard to be well addressed due to the limitation of collective influence in complex scenarios. These approaches model the interactions through the spatial relations between the target obstacle and its neighbors. However, they oversimplify the challenge since the training stage of the interactions lacks effective supervision. As a result, these models are far from promising. More intuitively, we transform the problem into calculating the interaction-aware drivable spaces and propose the CNN-based VisionNet for trajectory prediction. The VisionNet accepts a sequence of motion states, i.e., location, velocity, and acceleration, to estimate the future drivable spaces. The reified interactions significantly increase the interpretation ability of the VisionNet and refine the prediction. To further advance the performance, we propose an interactive loss to guide the generation of the drivable spaces. Experiments on multiple public datasets demonstrate the effectiveness of the proposed VisionNet.