VISTA: A Vision and Intent-Aware Social Attention Framework for Multi-Agent Trajectory Prediction
This addresses the need for safer autonomous systems by improving trajectory realism and reducing collisions in interactive scenarios, though it is incremental as it builds on existing goal-conditioned and transformer-based methods.
The paper tackled the problem of multi-agent trajectory prediction in dense environments by proposing VISTA, a recursive goal-conditioned transformer that integrates long-term intent and social interactions, resulting in state-of-the-art accuracy and a reduction in collision rates from 2.14% to 0.03% on the MADRAS benchmark and zero collisions on SDD.
Multi-agent trajectory prediction is crucial for autonomous systems operating in dense, interactive environments. Existing methods often fail to jointly capture agents' long-term goals and their fine-grained social interactions, which leads to unrealistic multi-agent futures. We propose VISTA, a recursive goal-conditioned transformer for multi-agent trajectory forecasting. VISTA combines (i) a cross-attention fusion module that integrates long-horizon intent with past motion, (ii) a social-token attention mechanism for flexible interaction modeling across agents, and (iii) pairwise attention maps that make social influence patterns interpretable at inference time. Our model turns single-agent goal-conditioned prediction into a coherent multi-agent forecasting framework. Beyond standard displacement metrics, we evaluate trajectory collision rates as a measure of joint realism. On the high-density MADRAS benchmark and on SDD, VISTA achieves state-of-the-art accuracy and substantially fewer collisions. On MADRAS, it reduces the average collision rate of strong baselines from 2.14 to 0.03 percent, and on SDD it attains zero collisions while improving ADE, FDE, and minFDE. These results show that VISTA generates socially compliant, goal-aware, and interpretable trajectories, making it promising for safety-critical autonomous systems.