With the gradual growth of road networks, the graphs used to represent road networks have been becoming larger and larger. It is a preferable choice for the user to outsource large-scale graphs for road networks to cloud and allow the cloud to store and query the information from graphs for him. It improves query efficiency greatly and reduces the user’s burden remarkably. To protect the privacy of graph data, the user has to encrypt the graph before uploading it to the cloud. It makes performing query on the encrypted graph become a challenge. In this paper, we propose the first scheme for performing a constrained top-$k$ nearest keyword query on an encrypted graph, even if a typo is made by the user. In the proposed scheme, we construct a new index structure based on the 2-hop cover label index to calculate the constrained shortest distance. In order to support fuzzy keyword query, we generate the fuzzy keyword set to build the keyword index in the graph. The security analysis shows the proposed scheme satisfies CQA-security and the experiments demonstrate the proposed scheme is efficient.

随着道路网络的逐渐壮大，用来表示道路网络的图形也越来越大。对于用户来说，将路网的大规模图外包给云，让云为他存储和查询图中的信息，是一个更好的选择。大大提高了查询效率，显着减轻了用户的负担。为了保护图数据的隐私，用户必须在将图上传到云端之前对其进行加密。它使得在加密图上执行查询成为一个挑战。在本文中，我们提出了第一个执行约束顶部的方案$克$加密图上的最近关键字查询，即使用户输入了拼写错误。在所提出的方案中，我们基于2-hop cover标签索引构造了一个新的索引结构来计算约束最短距离。为了支持模糊关键字查询，我们生成模糊关键字集来构建图中的关键字索引。安全性分析表明所提出的方案满足CQA-security，实验证明所提出的方案是有效的。