BucketKD: A Safety-Aware Bucket-Based Knowledge Distillation Framework for End-to-End Motion Planning
2026-07-12 • Robotics
RoboticsMachine Learning
AI summaryⓘ
The authors developed BucketKD, a method to make self-driving car software smaller and safer. Instead of simplifying the environment too much, they group important surroundings into buckets that keep more useful details without slowing things down. They also add a way for the model to pay more attention to risky spots, like close obstacles or potential collisions, to keep the car safer. Tests show their method is more accurate and safer while using less computing power compared to other approaches.
end-to-end motion planningknowledge distillationbucket-based discretizationwaypoint attention mechanismtime-to-collision (TTC)autonomous drivingdeep neural networksCARLA simulatorplanning accuracymodel compression
Authors
Md Nahidul Islam, Mohd Hasan Ali, Dipankar Dasgupta, Myounggyu Won
Abstract
End-to-end motion planning has emerged as a promising paradigm in autonomous driving, directly mapping raw sensor data to control commands via deep neural networks. Despite its advantages, its large model size hinders deployment in resource-constrained platforms. In this paper, we present BucketKD, a bucket-based knowledge distillation framework that yields compact and safety-aware end-to-end planners. Compared to the state-of-the-art approach, which relies on simplified planning state representations, BucketKD discretizes critical environmental variables into adaptive buckets that capture richer scene semantics while preserving efficiency. In addition, we design a safety-aware waypoint attention mechanism that evaluates each waypoint's risk level by accounting for both obstacle proximity and relative motion through a time-to-collision (TTC) formulation widely used in transportation research. This enables the student model to better retain safety-critical behaviors during distillation. Extensive experiments in CARLA using the Bench2Drive dataset show that BucketKD significantly outperforms the state-of-the-art in both planning accuracy and safety while maintaining strong compression ratios.