汽车工程学报2026，Vol.16Issue(2)：215-226,12.DOI:10.3969/j.issn.2095‒1469.2026.02.05
基于多结点仿真的自动驾驶场景数据保护方法
A Multi-Node Simulation Method for Autonomous Driving Scenarios Data Protection
摘要
Abstract
High-quality testing scenarios significantly enhance the efficiency of simulation testing for autonomous vehicles and possess substantial data value.To address the issue of data leakage caused by the full exposure of scenario data to simulation engines or platforms in existing simulation testing processes,this paper conducts an in-depth analysis of the OpenSCENARIO format standard,semantic content,and mapping relationships,proposing a novel scenario data protection method based on multi-node simulation.This method divides a scenario into multiple sub-segments and distributes them to different computing nodes for simulation.This enables fragmented operation of the complete scenario within a distributed simulation framework,effectively preventing full exposure of the original data.Furthermore,obfuscation techniques are introduced to preprocess the textual content of the sub-scenario data,further weakening the correlations among segments and increasing the difficulty of the reconstructing the complete scenario.Experimental results demonstrate that the simulation of these generated sub-scenario segments shows minimal deviation from the original simulation,successfully completing the simulation tasks while avoiding scenario data leakage.关键词
自动驾驶/仿真测试/OpenSCENARIO/数据保护Key words
autonomous driving/simulation test/OpenSCENARIO/data protection分类
信息技术与安全科学引用本文复制引用
彭海洋,计卫星,刘法旺..基于多结点仿真的自动驾驶场景数据保护方法[J].汽车工程学报,2026,16(2):215-226,12.基金项目
科技创新2030"新一代人工智能"重大项目(2022ZD0116311) （2022ZD0116311）
