用学习误差可控的模型，让自动驾驶更安全地应对周围车辆不确定性。
Interactive Trajectory Planning with Learning-based Distributionally Robust Model Predictive Control and Markov Systems
- 结合PAC学习与分布鲁棒优化，建模周围车辆决策不确定性
- 样本越多越接近理想规划，样本少时更保守稳健
- 适合自动驾驶等需兼顾安全与效率的交互式路径规划
我们研究在周围智能体决策不确定下的交互式轨迹规划问题。为控制自身代理（ego-agent），目标是首先学习其决策分布，并求解随机模型预测控制（SMPC）问题。为应对学习分布带来的误差，我们证明可通过概率近似正确（PAC）学习与分布鲁棒（DR）优化相结合，得到能考虑学习模型误差的解决方案。结果表明，基于PAC学习的分布鲁棒模型预测控制（DR-MPC）框架可依据可用样本数量，在鲁棒性极强的MPC与理想化的全能型SMPC之间进行插值，实现安全性与性能的动态平衡。
原文摘要 · Abstract (English)
We investigate interactive trajectory planning subject to uncertainty in the decisions of surrounding agents. To control the ego-agent, we aim to first learn the decision distribution and solve a Stochastic Model Predictive Control (SMPC) problem. To account for errors in the learned distribution, we show that it is possible to utilize Probably Approximately Correct (PAC) learning in combination with Distributionally Robust (DR) optimization to obtain a solution which accounts for the errors induced by the learning model. The results indicate that our PAC learning-based DR-MPC framework provides a method to interpolate between a robust MPC and an omnipotent SMPC, based on the available number of samples.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。