Latent Variable Sequential Set Transformers For Joint Multi-Agent Motion Prediction

TL;DR

This paper introduces AutoBots, a novel encoder-decoder architecture using latent variable sequential set transformers for efficient, scene-consistent multi-agent trajectory prediction, achieving top results on key benchmarks.

Contribution

The paper proposes AutoBots, a new model combining temporal and social attention modules for joint multi-agent motion prediction, capable of single-pass inference and trained efficiently on a single GPU.

Findings

Top results on nuScenes vehicle motion prediction leaderboard

Strong performance on Argoverse vehicle prediction challenge

Effective multi-agent social prediction on TrajNet++ dataset

Abstract

Robust multi-agent trajectory prediction is essential for the safe control of robotic systems. A major challenge is to efficiently learn a representation that approximates the true joint distribution of contextual, social, and temporal information to enable planning. We propose Latent Variable Sequential Set Transformers which are encoder-decoder architectures that generate scene-consistent multi-agent trajectories. We refer to these architectures as "AutoBots". The encoder is a stack of interleaved temporal and social multi-head self-attention (MHSA) modules which alternately perform equivariant processing across the temporal and social dimensions. The decoder employs learnable seed parameters in combination with temporal and social MHSA modules allowing it to perform inference over the entire future scene in a single forward pass efficiently. AutoBots can produce either the trajectory of one ego-agent or a distribution over the future trajectories for all agents in the scene. For the single-agent prediction case, our model achieves top results on the global nuScenes vehicle motion prediction leaderboard, and produces strong results on the Argoverse vehicle prediction challenge. In the multi-agent setting, we evaluate on the synthetic partition of TrajNet++ dataset to showcase the model's socially-consistent predictions. We also demonstrate our model on general sequences of sets and provide illustrative experiments modelling the sequential structure of the multiple strokes that make up symbols in the Omniglot data. A distinguishing feature of AutoBots is that all models are trainable on a single desktop GPU (1080 Ti) in under 48h.