부산대학교 도서관

Recently, real-time geometric data retrieval is the key technology for real-time mobility services to extract spatial-temporal data of moving objects (e.g. self-driving car, connected car and people using augmented reality) inside a static area (e.g. roads, residential area). Also, the high-definition map provides the high precise geographic polygon data on the static area. In general, the system of the real-time geometric data retrieval consists in relational database managing geographic polygon data, key-value store managing spatial-temporal data of the moving objects, and points-in-polygon algorithm for the extraction described in above. However, the system to apply high-definition map for geographic polygon data has two problems; First, too much vertices of high-definition map deteriorates the performance of the points-in-polygon algorithm. Secondly, a user cannot give an appropriate search query since the user does not know the exact shape of geographic polygon data stored in SDB beforehand, only knows the approximate location of it. To solve these two problems, this paper proposes two technologies; (1) reduction of vertices of high-definition map using Douglas-Peucker algorithm, and (2) PG-code, geo code for indexing of geographic polygon data. As the result, evaluation of the system implementing these proposals confirmed that our system can satisfy the requirements of real-time mobility services.