Guiding Neuro-Symbolic Scenario Generation with Spatio-Temporal Logic

2026-05-18

AI Summary

This work addresses the inefficiency of conventional autonomous driving safety testing, which relies heavily on extensive real-world road trials and struggles to effectively uncover rare yet critical safety-critical edge cases. The authors propose a novel approach that, for the first time, integrates differentiable signal temporal logic (STL) with a multi-agent trajectory diffusion model. By performing gradient-based optimization in the latent space, the method enables targeted generation of realistic traffic scenarios that simultaneously adhere to empirical data distributions and satisfy user-specified safety properties. This framework facilitates interpretable, efficient, and controllable synthesis of critical test scenarios, substantially enhancing the stress-testing capability of autonomous driving systems.

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 STRELGen, a scalable framework for the targeted generation of safety-critical driving scenarios. STRELGen 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 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. STRELGen thus provides a flexible, interpretable, and powerful tool for stress-testing autonomous driving systems, moving beyond the limitations of brute-force data collection.

Problem

• autonomous driving
• safety evaluation
• edge cases
• scenario generation
• Spatio-Temporal Logic

Innovation

• Spatio-Temporal Logic
• Diffusion Model
• Scenario Generation
• Autonomous Driving Safety
• Differentiable Monitoring