A Python implementation of the ST-Matching algorithm for low-sampling-rate GPS trajectories [1].
[1] Lou, Y., Zhang, C., Zheng, Y., Xie, X., Wang, W. and Huang, Y., 2009, November. Map-matching for low-sampling-rate GPS trajectories. In Proceedings of the 17th ACM SIGSPATIAL international conference on advances in geographic information systems (pp. 352-361). ACM.
Note that codes of the 'S (spatial)' part of the algorithm is currently unavailable in this distribution.
Files containing the geographical information of the road network you want to match your trajectories to are requied:
(1) node file: a comma-separated file cotaining at least three columns: ['node', 'lng', 'lat'], where 'node' is the identification of the breakdown points of your road network, 'lng' and 'lat' the longitudes and latitudes of nodes.
(2) edge file: a comma-separated file cotaining at least three columns: ['edge', 's_node', 'e_node', 's_lng', 's_lat', 'e_lng', 'e_lat', 'c_lng', 'c_lat'], where 'edge' is the identification of the road segments of your road network, 's_node' and 'e_node' the begining nodes and ending nodes of segments, 's_lng', 's_lat' the coordinates of 's_node', 'e_lng', 'e_lat' the coordinates of 'e_node', 'c_lng', 'c_lat' the coordinates of the mid points of edges.
(3) network properties file: a comma-separated file cotaining at least four columns: ['section_id', 's_node', 'e_node', 'length'], where 'section_id' is the identification of the road segments of your road network, 's_node' and 'e_node' the begining nodes and ending nodes of segments, 'length' the lengths of road segments.
The path of above three files should be specified in STmatching_distribution_ver.py.
The trajectory file must be a comma-separated file cotaining at least three columns: ['TRAJ_ID', 'LON', 'LAT'], where 'TRAJ_ID' is the identification of each trajectory.
An example to run this library in parallel:
from STmatching_distribution_ver import *
from multiprocessing import cpu_count, Pool
trajectory = pd.read_csv('your_trajectory_path')
trajectory = data_convert(trajectory)
pool = Pool()
match_results = pool.map(trajectory_matching, trajectory)
pool.close()
match_results = pd.concat(match_results, ignore_index=True)
Matching results are restored in a dataframe match_results: ['TRAJ_ID', 'MATCHED_EDGE', 'MATCHED_NODE'], where 'MATCHED_EDGE' is the matched sequence of road segments, and 'MATCHED_NODE' the matched sequence of road nodes.
ST-Matching is under Apache License. See LICENSE file for full license text.