Controllable Generative Trajectory Prediction via Weak Preference Alignment
This work addresses a domain-specific problem for autonomous vehicle planning by providing a cost-effective way to enhance generative models, though it is incremental as it builds on existing CVAE methods.
The paper tackles the problem of generating controllably diverse trajectories for autonomous vehicle planning, where existing methods lack control over diversity, and demonstrates that their proposed PrefCVAE framework enables controllable predictions without degrading baseline accuracy.
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.