Bias Behind the Wheel: Fairness Testing of Autonomous Driving Systems
It addresses fairness problems in pedestrian detection for autonomous driving, which is crucial for safety but under-explored, though it is incremental as it applies existing fairness testing methods to this domain.
This paper tackles fairness issues in autonomous driving systems by testing eight state-of-the-art pedestrian detectors across demographic groups, revealing significant biases such as a 20.14% higher undetection rate for children compared to adults and gender biases during nighttime.
This paper conducts fairness testing of automated pedestrian detection, a crucial but under-explored issue in autonomous driving systems. We evaluate eight state-of-the-art deep learning-based pedestrian detectors across demographic groups on large-scale real-world datasets. To enable thorough fairness testing, we provide extensive annotations for the datasets, resulting in 8,311 images with 16,070 gender labels, 20,115 age labels, and 3,513 skin tone labels. Our findings reveal significant fairness issues, particularly related to age. The proportion of undetected children is 20.14% higher compared to adults. Furthermore, we explore how various driving scenarios affect the fairness of pedestrian detectors. We find that pedestrian detectors demonstrate significant gender biases during night time, potentially exacerbating the prevalent societal issue of female safety concerns during nighttime out. Moreover, we observe that pedestrian detectors can demonstrate both enhanced fairness and superior performance under specific driving conditions, which challenges the fairness-performance trade-off theory widely acknowledged in the fairness literature. We publicly release the code, data, and results to support future research on fairness in autonomous driving.