Distribution shifts between operational domains can severely affect the performance of learned models in self-driving vehicles (SDVs). While this is a well-established problem, prior work has mostly explored naive solutions such as fine-tuning, focusing on the motion prediction task. In this work, we explore novel adaptation strategies for differentiable autonomy stacks consisting of prediction, planning, and control, perform evaluation in closed-loop, and investigate the often-overlooked issue of catastrophic forgetting. Specifically, we introduce two simple yet effective techniques: a low-rank residual decoder (LoRD) and multi-task fine-tuning. Through experiments across three models conducted on two real-world autonomous driving datasets (nuPlan, exiD), we demonstrate the effectiveness of our methods and highlight a significant performance gap between open-loop and closed-loop evaluation in prior approaches. Our approach improves forgetting by up to 23.33% and the closed-loop OOD driving score by 9.93% in comparison to standard fine-tuning.

翻译：运行域之间的分布偏移会严重影响自动驾驶车辆（SDV）中学习模型的性能。尽管这是一个公认的问题，但先前的研究主要探索了诸如微调之类的简单解决方案，并集中于运动预测任务。在本研究中，我们探索了由预测、规划与控制构成的可微分自动驾驶栈的新颖适应策略，在闭环中进行评估，并研究了常被忽视的灾难性遗忘问题。具体而言，我们引入了两种简单而有效的技术：低秩残差解码器（LoRD）和多任务微调。通过在两个真实世界自动驾驶数据集（nuPlan, exiD）上对三个模型进行的实验，我们证明了我们方法的有效性，并凸显了先前方法中开环与闭环评估之间的显著性能差距。与标准微调相比，我们的方法将遗忘降低了高达23.33%，并将闭环OOD驾驶分数提高了9.93%。