M2DA: Multi-Modal Fusion Transformer Incorporating Driver Attention for Autonomous Driving

TL;DR

This paper introduces M2DA, a multi-modal fusion transformer with driver attention that improves environment perception and scene understanding in autonomous driving, achieving state-of-the-art results in simulation.

Contribution

The paper proposes a novel Lidar-Vision-Attention Fusion module and incorporates driver attention to enhance multi-modal data integration and human-like scene understanding.

Findings

Achieves state-of-the-art performance on CARLA benchmarks

Requires less data for training compared to existing methods

Enhances safety by accurately identifying critical agents

Abstract

End-to-end autonomous driving has witnessed remarkable progress. However, the extensive deployment of autonomous vehicles has yet to be realized, primarily due to 1) inefficient multi-modal environment perception: how to integrate data from multi-modal sensors more efficiently; 2) non-human-like scene understanding: how to effectively locate and predict critical risky agents in traffic scenarios like an experienced driver. To overcome these challenges, in this paper, we propose a Multi-Modal fusion transformer incorporating Driver Attention (M2DA) for autonomous driving. To better fuse multi-modal data and achieve higher alignment between different modalities, a novel Lidar-Vision-Attention-based Fusion (LVAFusion) module is proposed. By incorporating driver attention, we empower the human-like scene understanding ability to autonomous vehicles to identify crucial areas within complex scenarios precisely and ensure safety. We conduct experiments on the CARLA simulator and achieve state-of-the-art performance with less data in closed-loop benchmarks. Source codes are available at https://anonymous.4open.science/r/M2DA-4772.