Accelerating the Evolution of Personalized Automated Lane Change through Lesson Learning

2024-05-13

arXiv.org

AI Summary

To address the challenges of online evolution, high computational overhead, and real-time adaptation to user preferences in personalized automated lane-change systems, this paper proposes a driver-intervention-driven curriculum learning framework. Methodologically, it integrates Gaussian Discriminant Analysis with apprenticeship learning to dynamically construct a safety-aware driving region, and combines online reward shaping with model predictive control for lightweight, adaptive trajectory policy optimization. Key contributions include: (i) the first introduction of an intervention-driven curriculum learning paradigm enabling experience-accumulated evolutionary adaptation; (ii) an average of only 13.8 iterations per evolutionary update and a response latency as low as 0.08 seconds; and (iii) empirical validation showing zero secondary interventions across all tests, with a 24% improvement in evolutionary efficiency over baseline methods.

Technology Category

• Search and Optimization: Learning to Search
• Machine Learning: Evolutionary Learning
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Studies of user behavior, including longitudinal effects of personalized systems

Abstract

Personalization is crucial for the widespread adoption of advanced driver assistance system. To match up with each user's preference, the online evolution capability is a must. However, conventional evolution methods learn from naturalistic driving data, which requires a lot computing power and cannot be applied online. To address this challenge, this paper proposes a lesson learning approach: learning from driver's takeover interventions. By leveraging online takeover data, the driving zone is generated to ensure perceived safety using Gaussian discriminant analysis. Real-time corrections to trajectory planning rewards are enacted through apprenticeship learning. Guided by the objective of optimizing rewards within the constraints of the driving zone, this approach employs model predictive control for trajectory planning. This lesson learning framework is highlighted for its faster evolution capability, adeptness at experience accumulating, assurance of perceived safety, and computational efficiency. Simulation results demonstrate that the proposed system consistently achieves a successful customization without further takeover interventions. Accumulated experience yields a 24% enhancement in evolution efficiency. The average number of learning iterations is only 13.8. The average computation time is 0.08 seconds.

Problem

• Enables online evolution of personalized lane change systems
• Reduces computing power needs via takeover intervention learning
• Ensures perceived safety with real-time trajectory corrections

Innovation

• Lesson learning from driver takeover interventions
• Gaussian discriminant analysis for safety zones
• Model predictive control for trajectory planning