改进的极小连通支配集SLAM数据关联方法

地图的极小连通支配集（MCDS）方法解决了机器人同时定位与地图创建（SLAM）过程中数据关联的规模随地图的不断增长而增加的问题。为了进一步优化MCDS方法的性能，对它进行了两处改进：一是延迟建立极小连通支配集；二是自适应地搜索极小连通支配集。K时刻的极小连通支配集子图延迟一个时间步而在K+1时刻建立，根据环境特征的疏密，搜索与K时刻接近的N个时间步内获得的地图数据，同时应用联合相容检验准则和分支定界搜索算法进行数据关联。仿真结果表明，改进的极小连通支配集方法的数据关联结果是可信的，大大降低了算法计算复杂度。

Data association method of SLAM based on improved minimal connected dominating set

The Minimal Connected Dominating Set (MCDS) method can solve the problem that the scale of data association increases with the map grows in process of Simultaneous Localization and Mapping-building (SLAM). Therefore, two improvements were introduced. Firstly, MCDS was constructed lingeringly. Secondly, MCDS was searched adaptively. The MCDS of K times was set up delay at K+1 times. According to the number of environment characteristics, the map data within N time was searched. At the same time, the joint compatibility criterion and branch and bound algorithm was used in data association. The simulation results indicate that data association results of minimizing connected dominating set are reliable, and the capability of reducing computational complexity is outstanding.