Multi-timescale feedback improves multi-agent traffic simulation consistency
Hi-FLoop: Hierarchical State-Feedback Loops for Multi-Timescale World Modeling
Computer Vision and Pattern RecognitionArtificial Intelligence
Summary
Simulating traffic with many vehicles over a long time is tricky because cars make decisions at different speeds and those decisions affect each other. The authors created a system called HI-FLOOP that breaks down the problem into layers working at different timescales, from broad goals to short-term control. This way, the simulated vehicles can coordinate better and adapt as the situation changes over time. Their method was tested on a public traffic dataset and showed improved accuracy in predicting vehicle behaviors.
What this means in practice
- •For autonomous vehicle developers: Simulate realistic multi-vehicle traffic scenarios with consistent interaction and multi-scale decision making for algorithm testing and training.
- •For urban traffic planners: Model and forecast diverse future traffic scenarios over multiple seconds to evaluate traffic flow under different control strategies.
Authors
Rx Fan, Zhan H
Abstract
Multi-agent traffic simulation seeks diverse, coordinated, and physically realistic futures from maps and observed history. Long-horizon closed-loop generation must reconcile multiple decision time scales while its context evolves with generated states. Existing methods often unfold long futures from the initial scene and resolve intent, interaction, and motion monolithically, weakening cross-scale consistency and adaptation. We present HI-FLOOP, a branch-consistent multi-timescale state-feedback framework. Eight scene-level Worlds represent joint hypotheses, and all agents share the selected World identity throughout an 8-second rollout. Within the branch, an 8-second Goal anchors intent, a 2-second Preview coordinates interactions, and 1-second Control produces physical motion. Every 0.5-second commit feeds back only its executed prefix as new facts, while unexecuted hypotheses never enter factual memory. Joint Preview Interaction (JPI) induces a sparse directed future graph from Preview and uses conflict probabilities and signed arrival-time differences to gate interaction refinement. For generated-state recovery, a prefix-frozen A-to-B cascade lets frozen Model A generate 0-1 seconds, then transfers typed physical state, admissible context, and the branch index, but no latent state, to an independent Model B for re-encoding and 1-2-second recovery. On the full H-D public-validation split of 955 scenarios, one complete S1 run yields an Overall score of 0.689987 with the official evaluator. Under agent-centric oracle evaluation, HI-FLOOP achieves oracle-minADE@8 of 1.196636 m over the 8-second horizon and 0.526 m over the 6-second horizon.