{"title":"CPT: A Configurable Predictability Testbed for DNN Inference in AVs","authors":"Liangkai Liu;Yanzhi Wang;Weisong Shi","doi":"10.26599/TST.2024.9010037","DOIUrl":null,"url":null,"abstract":"Predictability is an essential challenge for autonomous vehicles (AVs)‘ safety. Deep neural networks have been widely deployed in the AV's perception pipeline. However, it is still an open question on how to guarantee the perception predictability for AV because there are millions of deep neural networks (DNNs) model combinations and system configurations when deploying DNNs in AVs. This paper proposes configurable predictability testbed (CPT), a configurable testbed for quantifying the predictability in AV's perception pipeline. CPT provides flexible configurations of the perception pipeline on data, DNN models, fusion policy, scheduling policies, and predictability metrics. On top of CPT, the researchers can profile and optimize the predictability issue caused by different application and system configurations. CPT has been open-sourced at: https://github.com/Torreskai0722/CPT.","PeriodicalId":48690,"journal":{"name":"Tsinghua Science and Technology","volume":"30 1","pages":"87-99"},"PeriodicalIF":6.6000,"publicationDate":"2024-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10676407","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Tsinghua Science and Technology","FirstCategoryId":"94","ListUrlMain":"https://ieeexplore.ieee.org/document/10676407/","RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"Multidisciplinary","Score":null,"Total":0}
引用次数: 0
Abstract
Predictability is an essential challenge for autonomous vehicles (AVs)‘ safety. Deep neural networks have been widely deployed in the AV's perception pipeline. However, it is still an open question on how to guarantee the perception predictability for AV because there are millions of deep neural networks (DNNs) model combinations and system configurations when deploying DNNs in AVs. This paper proposes configurable predictability testbed (CPT), a configurable testbed for quantifying the predictability in AV's perception pipeline. CPT provides flexible configurations of the perception pipeline on data, DNN models, fusion policy, scheduling policies, and predictability metrics. On top of CPT, the researchers can profile and optimize the predictability issue caused by different application and system configurations. CPT has been open-sourced at: https://github.com/Torreskai0722/CPT.
可预测性是自动驾驶汽车(AV)安全性面临的一项重要挑战。深度神经网络已被广泛应用于自动驾驶汽车的感知管道。然而,由于在 AV 中部署深度神经网络时存在数百万种深度神经网络(DNN)模型组合和系统配置,因此如何保证 AV 的感知可预测性仍是一个未决问题。本文提出了可配置可预测性测试平台(CPT),这是一种用于量化 AV 感知管道可预测性的可配置测试平台。CPT 对感知管道的数据、DNN 模型、融合策略、调度策略和可预测性指标进行了灵活配置。在 CPT 的基础上,研究人员可以剖析和优化由不同应用和系统配置引起的可预测性问题。CPT 已开源:https://github.com/Torreskai0722/CPT。
期刊介绍:
Tsinghua Science and Technology (Tsinghua Sci Technol) started publication in 1996. It is an international academic journal sponsored by Tsinghua University and is published bimonthly. This journal aims at presenting the up-to-date scientific achievements in computer science, electronic engineering, and other IT fields. Contributions all over the world are welcome.