哈哈哈sorry，弃坑了，在OSM导入到PostgreSQL这一步应该是有问题的，导入的table缺少一些字段，我不晓得原作者的导入方法是什么，以后有兴趣再研究哈哈哈！大家有兴趣的也可以尝试下~

ST-Matching is proposed in Map-matching for low-sampling-rate GPS trajectories, which is a map matching algorithm usually used for handling low sampling rate GPS trajectories. This paper is also avaliable in my repository, you can download it.

Then, I will introduce the flow of ST-Matching and how to run this program.

In this section, I will show a flow chart of ST-Matching algorithm. To be added...

• PostgreSQL 10
  • NOTE : How to operate PostgreSQL using C/C++?
  • The thing needs to know is that if you install a x64 version, you should compile the project in x64, otherwise it would be ERROR.
• Visual Studio 2017
  • recomended, some other IDEs are OK
• PostGIS 2.4.3
• osm2pgsql: OpenStreetMap data to PostgreSQL converter 0.92.0
  • Windows users should NOTICE that, Cross-compilation problems with osm2pgsql are documented on github. Whilst these problems are being worked on, an alternative compilation path for windows users targeting Cygwin has been used. Installation for Windows is here

Before running the program, we should import map file *.osm into database. So we can utilize osm2pgsql to do this. Before running osm2pgsql, you first MUST create your postgresql database and enable postgis on it. Then, you can import through command as following:

osm2pgsql -s -U postgres -d nameofdatabase /file/path/toosm/fileorpbf/name.osm -W

which U denotes your username, d denotes database name, or you can get more details through osm2pgsql -help. Then, enter your database's password, there will be some screen outputs like this:

Using XML parser.
Processing: Node(899k 179.9k/s) Way(124k 62.48k/s) Relation(1560 1560.00/s)  parse time: 8s
Node stats: total(899390), max(5463903487) in 5s
Way stats: total(124957), max(567723766) in 2s
Relation stats: total(1569), max(8055906) in 1s
...
...
node cache: stored: 899390(100.00%), storage efficiency: 50.00% (dense blocks: 0, sparse nodes: 899390), hit rate: 100.01%

Osm2pgsql took 20s overall

It's done.

You need to modify the config.ini in the project folder according to your database configuration, for example:

#ST-Matching配置文件,为方便处理，每一行“=”左边为配置的属性，“=”右边请空一格写入属性值，“=”左右两边请留一个空格。每行属性上一行是对其作用的解释，以“#”作为解释的开头。

#数据库名称
dbname = shanghai_osm
#数据库端口号
dbport = 5432
#数据库地址
dbaddr = 127.0.0.1
#道路表名称
roadTN = road_network
#线程数，声明使用多少线程来进行计算
threadnum = 3
#每个轨迹点最多选多少个候选点。建议为5个
K = 5
#求最优路径时选取当前值最大的top个点用作下一个点的计算
top = 3
#选取某轨迹点的候选点时，距离轨迹点的范围，单位为Km
R = 0.1
#正态分布的方差，单位为Km
sigma = 0.02

Usually, modifing the first four items is enough if you have no idea about the other five items.

The content of input gps file should be formatted as following, which respectively denote id,timestamp,longitude,latitude in order:

13301128438,2013-10-16 23:50:24,116.366653,39.938328
13301128438,2013-10-16 23:50:48,116.366623,39.937351
13301128438,2013-10-16 23:51:48,116.366722,39.933094
13301128438,2013-10-16 23:54:24,116.367302,39.921505
...

We should notice that the id segments are the same in a file, because the ST-Matching algorihm can only match one trajectory at a time.

This project is based on the same name repository of @AzoresCabbage. I try to reconstruct the project and make it easier to use. If there is any problem, please let me know.