WOMD-Reasoning: A Large-Scale Dataset and Benchmark for Interaction and Intention Reasoning in Driving

2024-07-05

AI Summary

Existing driving scene understanding lacks large-scale, multimodal language datasets explicitly designed for traffic-rule-compliant and human-intention-driven long-range interaction reasoning. Method: We introduce WOMD-Reasoning—the first large-scale multimodal driving question-answering dataset (3M Q&A pairs), incorporating high-definition maps, agent trajectories, and bird’s-eye-view (BEV) images, and systematically modeling rule- and intention-triggered long-distance interactions. We further propose Motion-LLaVA, a motion-language joint reasoning architecture trained via instruction tuning and chain-of-thought prompting. Contribution/Results: Motion-LLaVA achieves significant improvements over state-of-the-art baselines on causal reasoning and intention recognition benchmarks. The dataset, code, and model are fully open-sourced, establishing a new benchmark for cognitive intelligence in autonomous driving.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Computer Vision: Multi-modal Vision
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

We propose Waymo Open Motion Dataset-Reasoning (WOMD-Reasoning), a comprehensive large-scale dataset with 3 million Q&As built on WOMD focusing on describing and reasoning interactions and intentions in driving scenarios. Existing language datasets for driving primarily capture interactions caused by close distances. However, interactions induced by traffic rules and human intentions, which can occur over long distances, are yet sufficiently covered. To address this, WOMD-Reasoning presents by far the largest multi-modal Q&A dataset on real-world driving scenarios, covering a wide range of driving topics from map descriptions and motion status descriptions to narratives and analyses of agents' interactions, behaviors, and intentions. We further introduce Motion-LLaVA, a motion-language model fine-tuned on the proposed dataset with robust interaction reasoning capabilities. We benchmark its performance across various configurations including different input modalities, reasoning techniques, and network architectures. The robust, diverse, and multi-modal nature of WOMD-Reasoning highlights its potential to advance future autonomous driving research and enable a broad range of applications. The dataset and its vision modal extension are available at https://waymo.com/open/download, and the codes&prompts to build it are available at https://github.com/yhli123/WOMD-Reasoning.

Problem

• Lack of dedicated language datasets for traffic rule-based interaction analysis
• Need for large-scale multimodal Q&A dataset on driving scenarios
• Enhancing interaction reasoning and traffic rule compliance in autonomous driving

Innovation

• Large-scale Q&A dataset for driving interactions
• Multi-modal dataset with 3 million Q&As
• Motion-language model fine-tuned for reasoning