Guiding Neuro-Symbolic Scenario Generation with Spatio-Temporal Logic

Lorenzo Bonin, Francesco Giacomarra, Luca Bortolussi, Jyotirmoy V. Deshmukh, Francesca Cairoli

Keywords: Scenario Generation, Autonomous Driving, Deep Generative Models, Spatio-Temporal Logic

TL;DR: Guiding diffusion-based models for multi-trajectories AD scenarios generation using STREL formulae

Abstract:

The rapid advancement of autonomous driving (AD) technologies has outpaced the development of robust safety evaluation methods.
Conventional testing relies on exposing AD systems to vast numbers of real-world traffic scenes---a brute-force approach that is prohibitively expensive and statistically ineffective at capturing the rare, safety-critical edge cases essential for validating real-world robustness.
To address this fundamental limitation, we introduce SCENGen, a scalable framework for the targeted generation of safety-critical driving scenarios. \ourmethod\ synergistically combines a multi-agent trajectory-generation diffusion model (DM) with Spatio-Temporal Logic (STREL) specifications that encode complex safety and realism properties through a highly interpretable formalism. Crucially, monitoring satisfaction levels of these specifications is differentiable, enabling gradient-based search. At inference time, we optimize directly over the DM's latent space to maximize STREL formula satisfaction.
The result is efficient generation of highly plausible yet safety-critical multi-agent scenarios that lie within the learned data distribution. SCENGen thus provides a flexible, interpretable, and powerful tool for stress-testing autonomous driving systems, moving beyond the limitations of brute-force data collection.

Area: Representation and Reasoning (RR)

Generative A I: I acknowledge that I have read and will follow this policy.

Submission Number: 1682