Motor vehicle path fitting algorithm based on low sampling data

Abstract

The invention discloses a motor vehicle path fitting algorithm based on low sampling data. The method comprises steps of S1, reading the configuration item information, and obtaining track data and road network data which need to be processed; S4, filtering the road network data, determining a road type field needing to be reserved according to the actual situation of a city, and storing each roadaccording to the geographic information; S7, for the whole track, constructing a migration probability matrix for adjacent track points, calculating the comprehensive probability between adjacent projection points, and finally obtaining a probability matrix between projection points corresponding to each track point in the whole track; and S8, obtaining a relation graph of the whole track according to the comprehensive probability matrix of the whole track calculated in the S7, and sequentially selecting the projection point with the maximum value as the optimal fitting path of the track. Themethod is advantaged in that the information such as spatial distribution and time consumption is comprehensively considered, the migration probability between adjacent mapping points is calculated on the basis of space analysis at the moment, local optimal probability path selection is carried out, and progressive high-precision path fitting is achieved.

Description

technical field [0001] The invention belongs to the technical field of urban traffic, and in particular relates to a motor vehicle path fitting algorithm based on low sampling data. Background technique [0002] Matching raw signaling data to a digital map or digital road network is often referred to as path fitting. Signaling data has problems such as low sampling rate, poor precision, and irregular ping-pong switching, so it cannot be applied to traditional path fitting algorithms. Vehicle path fitting is an essential preprocessing step for many applications such as traffic flow analysis, driving route planning, etc. At present, most path fitting algorithms use local or incremental algorithms, which completely ignore the correlation between adjacent points, and are only suitable for GPS data with a high sampling rate. For signaling data with a low sampling rate, the processing results are poor . To this end, we propose a motor vehicle path fitting algorithm based on low...

AI Technical Summary

Problems solved by technology

At present, most path fitting algorithms use local or incremental algorithms, which completely ignore the correlation between adjacent points, and are only suitable for GPS data with a high sampling rate. For signaling data with a low sampling rate, the processing results are poor

Images