arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2412.03792cs.ROcs.AIcs.SYeess.SY

感知不确定性下的安全自适应巡航控制:一种深度集成与共形管模型预测控制方法

Safe Adaptive Cruise Control Under Perception Uncertainty: A Deep Ensemble and Conformal Tube Model Predictive Control Approach

University of Michigan(密歇根大学)

Xiao Li, Anouck Girard, Ilya Kolmanovsky

AI总结:

本文融合Deep Ensemble与Conformal Prediction量化RGB感知不确定性,并设计Conformal Tube MPC自适应巡航控制器,在高保真仿真及OOD场景中实现概率安全的速度跟踪与车距保持。

摘要:

自动驾驶严重依赖感知系统来解释环境以进行决策。为增强这些安全关键应用中的鲁棒性,本文考虑将深度神经网络(DNN)回归器的深度集成(Deep Ensemble)与共形预测(Conformal Prediction)相结合,以预测并量化不确定性。在自适应巡航控制(Adaptive Cruise Control)场景中,所提方法根据RGB图像进行状态与不确定性估计,将DNN感知不确定性告知下游控制器。设计了采用共形管模型预测控制(Conformal Tube Model Predictive Control)的自适应巡航控制器,以确保概率安全性。高保真仿真器评估表明,该算法在速度跟踪和安全距离保持方面有效,包括在分布外(Out-Of-Distribution)场景中。

英文摘要:

Autonomous driving heavily relies on perception systems to interpret the environment for decision-making. To enhance robustness in these safety critical applications, this paper considers a Deep Ensemble of Deep Neural Network regressors integrated with Conformal Prediction to predict and quantify uncertainties. In the Adaptive Cruise Control setting, the proposed method performs state and uncertainty estimation from RGB images, informing the downstream controller of the DNN perception uncertainties. An adaptive cruise controller using Conformal Tube Model Predictive Control is designed to ensure probabilistic safety. Evaluations with a high-fidelity simulator demonstrate the algorithm's effectiveness in speed tracking and safe distance maintaining, including in Out-Of-Distribution scenarios.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑