Extended Field of View Analysis for VideoGAN-based Trajectory Generation

2026-08-03

AI Summary

This work addresses the challenge of generating realistic and diverse vehicle trajectories in complex traffic scenarios by proposing a VideoGAN-based generative framework. The approach effectively models large-scale dynamic traffic environments through semantic bird’s-eye-view representations, a graph-structured trajectory association mechanism, and a field-of-view expansion strategy. A novel quantitative evaluation metric is introduced to assess object persistence and hallucination phenomena in generated videos. Trained within 150 GPU hours, the model achieves inference times under 20 ms for 20-second scenes while producing trajectories that exhibit high fidelity in both statistical properties and spatial relationships. The framework thus offers an efficient solution to support prediction, planning, and simulation requirements for high-level autonomous driving systems.

Abstract

Realistic and diverse trajectory generation is central to enabling higher levels of vehicle automation. While rule-based and classical learning-based methods may struggle to capture the complexity of traffic behavior, generative models have already demonstrated in other fields that they can handle a comparable level of complexity. In this paper, we build upon previous work on generative adversarial network (GAN)-based semantic bird's-eye-view traffic generation and extend the proposed framework in several key aspects. We improve the semantic representation, replace the trajectory extraction procedure with a graph-based association method, and systematically investigate increasingly larger fields of view. In addition, we introduce a quantitative evaluation framework to assess hallucinations and object permanence in generated videos. Our experiments demonstrate that the framework generalizes to larger and more complex traffic scenes while maintaining statistically realistic trajectories and coherent spatial relationships between traffic participants. Within 150GPU hours of training and with inference times below 20ms for scenes of up to 20s, our results demonstrate that video-based GANs remain an efficient and scalable approach for realistic trajectory generation, even in substantially larger traffic scenes, making them well suited for downstream tasks such as prediction, planning, and simulation in automated driving.

Problem

• trajectory generation
• vehicle automation
• field of view
• traffic scenes
• realistic trajectories

Innovation

• VideoGAN
• trajectory generation
• field of view
• graph-based association
• object permanence