AG-GAN: An Attentive Group-Aware GAN for pedestrian trajectory prediction

Understanding human behaviors in crowded scenarios requires analyzing not only the position of the subjects in space, but also the scene context. Existing approaches mostly rely on the motion history of each pedestrian and model the interactions among people by considering the entire surrounding neighborhood. In our approach, we address the problem of motion prediction by applying coherent group clustering and a global attention mechanism on the LSTM-based Generative Adversarial Networks (GANs). The proposed model consists of an attentive group-aware GAN that observes the agents' past motion and predicts future paths, using (i) a group pooling module to model neighborhood interaction, and (ii) an attention module to specifically focus on hidden states. The experimental results demonstrate that our proposal outperforms state-of-the-art models on common benchmark datasets, and is able to generate socially-acceptable trajectories.

Automated summary

Understanding human behaviors in crowded scenarios requires analyzing both the position of the subjects in space and the scene context. Existing approaches rely on motion history and interactions among people, while the proposed model uses coherent group clustering and a global attention mechanism to address motion prediction. The attentive group-aware GAN outperformed state-of-the-art models on benchmark datasets and generated socially-acceptable trajectories.