Misbehavior Forecasting for Focused Autonomous Driving Systems Testing

2025-12-21

AI Summary

To address the low efficiency and high cost of failure detection in autonomous driving simulation testing, this paper proposes a proactive near-miss-driven fuzzing method. First, a state-evolution prediction model is constructed to anticipate potential near-miss states of the ego vehicle. Subsequently, localized gradient-aware perturbations are applied within the identified near-miss neighborhoods to conduct targeted fuzzing. This approach introduces a novel closed-loop “prediction–identification–perturbation” mechanism, enabling early, precise, and efficient discovery of erroneous behaviors. Experimental results demonstrate that, compared to random testing and the state-of-the-art failure predictor, our method improves failure detection rates by 128.7% and 38.1%, respectively, while accelerating detection speed by 2.49× and 1.42×. When coordinated with DriveFuzz, the joint detection capability increases by 93.9%.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Intelligent Robots: State Estimation
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

Simulation-based testing is the standard practice for assessing the reliability of self-driving cars' software before deployment. Existing bug-finding techniques are either unreliable or expensive. We build on the insight that near misses observed during simulations may point to potential failures. We propose Foresee, a technique that identifies near misses using a misbehavior forecaster that computes possible future states of the ego-vehicle under test. Foresee performs local fuzzing in the neighborhood of each candidate near miss to surface previously unknown failures. In our empirical study, we evaluate the effectiveness of different configurations of Foresee using several scenarios provided in the CARLA simulator on both end-to-end and modular self-driving systems and examine its complementarity with the state-of-the-art fuzzer DriveFuzz. Our results show that Foresee is both more effective and more efficient than the baselines. Foresee exposes 128.70% and 38.09% more failures than a random approach and a state-of-the-art failure predictor while being 2.49x and 1.42x faster, respectively. Moreover, when used in combination with DriveFuzz, Foresee enhances failure detection by up to 93.94%.

Problem

• Forecasting near misses to predict potential autonomous driving failures
• Enhancing simulation-based testing efficiency for self-driving software
• Improving failure detection through local fuzzing around identified near misses

Innovation

• Forecasts near misses using future state computation
• Performs local fuzzing around candidate near misses
• Enhances failure detection when combined with DriveFuzz