Page 188 - 《武汉大学学报(信息科学版)》2025年第10期
P. 188
第 50 卷第 10 期 黄 龙等:智能网联汽车高精地图安全监管分析与技术研究 2123
ly in the stage of high-level automated driving above L3 level, high-definition maps can greatly reduce the
requirements for the arithmetic power of the vehicles, and help solve the problem of realizing better auto‑
mated driving under the environment of the high arithmetic power hardware “necklace”. However, the
contradiction between the urgent demand for high-definition maps for automated driving and the safety risks
in practical applications is becoming more and more prominent. Methods: Based on the urgent demand for
high-definition maps for automated driving and the safety risks in applications, we focus on safety supervivion
of high-definition map for intelligent connected vehicle and aim to provide effective solution ideas for the
safe application. First, the current situation of high-definition map safety supervision at home and abroad is
introduced. Based on the systematic analysis of the safety risks throughout the whole lifecycle of high-defi‑
nition map, from data collection, production, processing, release and application, a high-definition map
safety supervision framework is proposed. Then, the geographic information decryption and desensitiza‑
tion, map safety scrutiny, spatial-temporal data provenance and other key technologies involved in the
framework are further sorted out. Results: Taking the high-definition map safety supervision pilot in Bei‑
jing as a case study, we introduce the current status of its application and exploration experience. After a se‑
ries of exploration, Beijing has established an advanced mechanism for dynamic supervision of the safety of
high-definition maps for intelligent connected vehicles by pre-registration, mid-monitoring, and post-tracing.
A platform for the safety supervision of maps is also set up, which comprehensively utilizes geographic in‑
formation, big data, 5G, artificial intelligence and other technology. Conclusions: High-definition maps,
as important basic resources and data elements, are an important link in the automated driving industry and
the spatial-temporal information industry chain. To better secure high-definition maps, on the one hand,
the government should empower policy propaganda and also solidify the supervision, on the other hand, en‑
terprises and institutions engaged in surveying and mapping activities should also strengthen safety aware‑
ness and protection.
Key words: high-definition map; safety supervision; geographic information safety; spatial-temporal da‑
ta; artificial intelligence
高精地图作为自动驾驶技术的核心支撑,发 过程不加以严格的监督与管控,可能对国家安全
挥着至关重要的作用。尽管近年来无图大模型 和社会稳定造成严重威胁。
的兴起为自动驾驶提供了新的技术路径,但由于 面对这一危机,行业整体亟需加强高精地图
车载计算硬件的高算力发展仍面临诸多挑战,高 产品及其采集、生产加工、应用过程的安全监管。
精地图凭借其数据的高精度、高维度、高丰富度 为此,近年来相关部门对高精地图的安全应用问
和高鲜度特性,在辅助车辆定位、增强环境感知 题的关注逐渐加深,同时行业各方已呈现维护高
以及实现精准路径规划和决策等方面依然不可 精地图数据安全的积极趋势,发展了一系列有针
或 缺 [1-2] 。 尤 其 是 在 道 路 标 线 磨 损 、交 通 设 施 遮 对性的安全监管技术,包括地理围栏、数字水印、
挡、大雪雾霾天气等场景下,高精地图相较纯视 数 据 加 密 等 ,旨 在 最 大 限 度 地 减 少 数 据 在 全 流
[4]
觉 、无图大模型 等方案具有明显优势。 程、全生命周期的安全风险。此外,区块链等新
[3]
然而,技术的飞速发展往往伴随着政策和监 兴技术的应用也提升了数据的透明度和可追溯
管的不足。一方面,若高精地图产品的质量与可 性。然而,目前学术与产业界尚缺少针对高精地
靠性良莠不齐,如精度、一致性、完整性等方面, 图安全审查技术的深入研究,缺乏对这些关键技
可能造成公共交通安全问题。在现实场景中,高 术的系统分析、总结和未来展望,这使得行业在
精地图的质量缺陷已在一定程度上受到消费市 制定有效的安全监管策略时面临挑战。
场 的 限 制 。 另 一 方 面 ,在 高 精 地 图 的 生 产 过 程 在此背景下,本文将探讨高精地图安全应用
中,涉及的数据处理手段复杂多样,数据在各个 全流程的安全风险,侧重于地图生产流程中涉及
处理方之间的流转也各具特性。这种复杂性显 国家地理信息的数据安全性监管,提出覆盖数据
著增加了敏感时空信息暴露的风险,包括数据泄 采集、数据生产加工、数据发布应用环节的安全
露、篡改或不当使用等行为。若对高精地图生产 隐患,为监管框架的设计与实施提供参考,并以

