An Improved Interactive-voting based Map Matching Algorithm Considering Path Correlation.

Map matching is a technology that aligns users' GPS position sequence with the road network on a digital map. Under low-sampling-rate conditions, existing interactive voting-based map matching algorithm leads to mismatching and low matching efficiency. Considering such problems, this paper proposes an improved interactive voting-based map matching algorithm considering path correlation by optimizing the observation probability and transition probability formulas to improve spatiotemporal analysis. Average speed and sampling time are used to estimate the path length and analyze the correlation between the estimated path and the actual path in order to reduce mismatching and improve the accuracy of matching. Utilizing three constraint conditions to filter erroneous candidate road segments improves accuracy and reduces matching time. The experimental results show that under various road conditions, the improved algorithm outperforms the compared algorithms. The matching accuracy can be maintained at over 90%, and the matching time is reduced by about 1ms compared to the comparison algorithms. [ABSTRACT FROM AUTHOR]