2026
CAPS: Context-Aware Priority Sampling for Enhanced Imitation Learning in Autonomous Driving
Hamidreza Mirkhani, Behzad Khamidehi, Ehsan Ahmadi, Mohammed Elmahgiubi, Weize Zhang, Fazel Arasteh +3
ICRA 2026poster
In this paper, we introduce Context-Aware Priority Sampling (CAPS), a novel method designed to enhance data efficiency in learning-based autonomous driving systems. CAPS addresses the challenge of imbalanced datasets in imitation learning by leveraging Vector Quantized Variational Autoencoders (VQ-V…