Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Validity Learning on Failures: Mitigating the Distribution Shift in Autonomous Vehicle Planning

Fazel Arasteh, Mohammed Elmahgiubi, Behzad Khamidehi, Hamidreza Mirkhani, Weize Zhang, Cao Tongtong, Kasra Rezaee

TL;DR

Validity Learning on Failures, VL(on failure), is proposed, which aims to discern valid trajectories within the current environmental context and outperforms the state-of-the-art methods by a large margin.

Abstract

The planning problem constitutes a fundamental aspect of the autonomous driving framework. Recent strides in representation learning have empowered vehicles to comprehend their surrounding environments, thereby facilitating the integration of learning-based planning strategies. Among these approaches, Imitation Learning stands out due to its notable training efficiency. However, traditional Imitation Learning methodologies encounter challenges associated with the co-variate shift phenomenon. We propose Validity Learning on Failures, VL(on failure), as a remedy to address this issue. The essence of our method lies in deploying a pre-trained planner across diverse scenarios. Instances where the planner deviates from its immediate objectives, such as maintaining a safe distance from obstacles or adhering to traffic rules, are flagged as failures. The states corresponding to these failures are compiled into a new dataset, termed the failure dataset. Notably, the absence of expert annotations for this data precludes the applicability of standard imitation learning approaches. To facilitate learning from the closed-loop mistakes, we introduce the VL objective which aims to discern valid trajectories within the current environmental context. Experimental evaluations conducted on both reactive CARLA simulation and non-reactive log-replay simulations reveal substantial enhancements in closed-loop metrics such as \textit{Score, Progress}, and Success Rate, underscoring the effectiveness of the proposed methodology. Further evaluations against the Bench2Drive benchmark demonstrate that VL(on failure) outperforms the state-of-the-art methods by a large margin.

Validity Learning on Failures: Mitigating the Distribution Shift in Autonomous Vehicle Planning

TL;DR

Validity Learning on Failures, VL(on failure), is proposed, which aims to discern valid trajectories within the current environmental context and outperforms the state-of-the-art methods by a large margin.

Abstract

The planning problem constitutes a fundamental aspect of the autonomous driving framework. Recent strides in representation learning have empowered vehicles to comprehend their surrounding environments, thereby facilitating the integration of learning-based planning strategies. Among these approaches, Imitation Learning stands out due to its notable training efficiency. However, traditional Imitation Learning methodologies encounter challenges associated with the co-variate shift phenomenon. We propose Validity Learning on Failures, VL(on failure), as a remedy to address this issue. The essence of our method lies in deploying a pre-trained planner across diverse scenarios. Instances where the planner deviates from its immediate objectives, such as maintaining a safe distance from obstacles or adhering to traffic rules, are flagged as failures. The states corresponding to these failures are compiled into a new dataset, termed the failure dataset. Notably, the absence of expert annotations for this data precludes the applicability of standard imitation learning approaches. To facilitate learning from the closed-loop mistakes, we introduce the VL objective which aims to discern valid trajectories within the current environmental context. Experimental evaluations conducted on both reactive CARLA simulation and non-reactive log-replay simulations reveal substantial enhancements in closed-loop metrics such as \textit{Score, Progress}, and Success Rate, underscoring the effectiveness of the proposed methodology. Further evaluations against the Bench2Drive benchmark demonstrate that VL(on failure) outperforms the state-of-the-art methods by a large margin.
Paper Structure (11 sections, 3 equations, 4 figures, 2 tables)

This paper contains 11 sections, 3 equations, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Imitation Learning (IL) with Sample-based Planner
  4. Failure States Data Collection
  5. Validity Learning (VL) on Failure States
  6. Evaluation
  7. Experiments Setup
  8. Experimental Results
  9. Ablation Study
  10. Limitations and Discussions
  11. Conclusion

Figures (4)

  • Figure 1: (Left) Imitation Learning (IL): increase the probability of the candidate trajectory closest to the expert's trajectory (green), (Right) Validity Learning (VL): increase the probability of the valid candidate trajectories (green)
  • Figure 2: (Top): Imitation Learning (IL) on Expert Labeled States, (Middle): Failure States Data Collection, (Bottom): Validity Learning (VL) on Unlabeled Failure States
  • Figure 3: VL(on failure) vs IL+RL: Reward over Training Scenarios (Moving Average, and 95% confidence intervals)
  • Figure 4: Effect of validity loss on imitation loss