Connectivity and Principal Component Analysis Based Network Decomposition for the Vehicle Navigation Path Analysis

It is necessary to decompose a large map into small pieces while performing vehicle navigation path analysis in an onboard device. A principal component analysis based network decomposition is constructed to solve this problem. An index of connectivity is proposed in this method, by which the most correlative vertices relative to the destination vertices can be extracted from the complete road network to compose smaller sub-maps. Thus the computation complexity of path analysis on these maps is greatly reduced without much accuracy loss. This method is applied in the distribute computing between logistics monitoring center and in-vehicle devices. Experimental evaluation shows that this method can effectively improve the computation performance of onboard devices.