MMFN: Multi-Modal-Fusion-Net for End-to-End Driving

TL;DR

This paper introduces MMFN, a multi-modal fusion network that effectively integrates sensor data and HD map features for improved end-to-end autonomous driving performance.

Contribution

The paper proposes a novel method for extracting and utilizing vectorized HD map features and incorporates a new expert module considering multi-road rules, enhancing driving accuracy.

Findings

Achieves superior driving performance over existing methods.

Efficiently extracts useful map features to avoid misleading information.

Enhances model with multi-road rule consideration.

Abstract

Inspired by the fact that humans use diverse sensory organs to perceive the world, sensors with different modalities are deployed in end-to-end driving to obtain the global context of the 3D scene. In previous works, camera and LiDAR inputs are fused through transformers for better driving performance. These inputs are normally further interpreted as high-level map information to assist navigation tasks. Nevertheless, extracting useful information from the complex map input is challenging, for redundant information may mislead the agent and negatively affect driving performance. We propose a novel approach to efficiently extract features from vectorized High-Definition (HD) maps and utilize them in the end-to-end driving tasks. In addition, we design a new expert to further enhance the model performance by considering multi-road rules. Experimental results prove that both of the proposed improvements enable our agent to achieve superior performance compared with other methods.