Road Crosses High Locality Sorting for Navigation Route Planning

Fast route planning in embedded navigation systems is restricted by main memory size and can only use external-memory algorithms, so high locality for external-memory roadnet data is the key issue for fast route planning. A continues quantitative binary sorting method based on the k-d tree structure is designed to sort the crosses in road networks. It has high spatial locality and support different size of blocking and fast random access. Applications in the whole Chinese road network show that it can support high efficient data access for breadth-first search based route planning algorithms.