Controllable Generative Trajectory Prediction via Weak Preference Alignment

2025-10-12

AI Summary

Existing trajectory prediction models for autonomous driving achieve high accuracy but suffer from insufficient multimodal diversity and limited semantic controllability, hindering safe motion planning. To address this, we propose PrefCVAE, a preference-guided conditional variational autoencoder framework that incorporates weak preference supervision—leveraging coarse-grained human preferences over semantic trajectory attributes (e.g., average speed)—to impose interpretable, semantics-aware constraints on the latent space. This enables controllable generation of diverse, semantically meaningful trajectory hypotheses without requiring expensive fine-grained annotations, thereby significantly reducing labeling overhead. PrefCVAE preserves prediction accuracy while supporting explicit modulation of the output distribution along interpretable semantic axes (e.g., aggressive vs. conservative). Extensive experiments demonstrate its superior balance of diversity control, interpretability, and practical applicability compared to state-of-the-art methods.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Knowledge Representation and Reasoning: Preferences
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Deep generative models such as conditional variational autoencoders (CVAEs) have shown great promise for predicting trajectories of surrounding agents in autonomous vehicle planning. State-of-the-art models have achieved remarkable accuracy in such prediction tasks. Besides accuracy, diversity is also crucial for safe planning because human behaviors are inherently uncertain and multimodal. However, existing methods generally lack a scheme to generate controllably diverse trajectories, which is arguably more useful than randomly diversified trajectories, to the end of safe planning. To address this, we propose PrefCVAE, an augmented CVAE framework that uses weakly labeled preference pairs to imbue latent variables with semantic attributes. Using average velocity as an example attribute, we demonstrate that PrefCVAE enables controllable, semantically meaningful predictions without degrading baseline accuracy. Our results show the effectiveness of preference supervision as a cost-effective way to enhance sampling-based generative models.

Problem

• Enabling controllable diverse trajectory predictions for safe planning
• Addressing lack of semantic control in generative trajectory models
• Using weak preference alignment to imbue trajectories with meaningful attributes

Innovation

• Uses weakly labeled preference pairs for alignment
• Augments CVAE with semantic attribute control
• Enables controllable diverse trajectory predictions