Authors: Wencheng Han, Dongqian Guo, Cheng-Zhong Xu, Jianbing Shen

Published on: January 08, 2024

Impact Score: 8.45

Arxiv code: Arxiv:2401.03641

Summary

• What is new: The introduction of DME-Driver, which combines a vision language model for decision-making with a planning-oriented perception model for signal generation.
• Why this is important: The need for improved explainability in decision logic and enhanced accuracy in environmental perception in autonomous driving systems.
• What the research proposes: The DME-Driver system that uses a vision language model to emulate experienced human driver logic and a planning-oriented perception model for accurate control signal generation.
• Results: Achieved high-precision planning accuracy that reflects a logical thinking process, trained on a newly created autonomous driving dataset.

Technical Details

Technological frameworks used:

Models used: Vision language model for logical decision-making; Planning-oriented perception model for environmental sensing and control signal generation.

Data used: New dataset capturing a diversity of human driver behaviors and intentions.

Potential Impact

Automotive industry, specifically companies developing autonomous driving technologies, and related software sectors.

Want to implement this idea in a business?

We have generated a startup concept here: ClearPath AI.