Recently, Particle filter becomes the most popular approach in mobile robot localization and has been applied with great success to a variety of state estimation problems. In this paper, the particle filter is applied in position tracking and global localization. Moreover, the posterior distribution of robot pose in global localization is usually multimodal due to the symmetry of the environment and ambiguous detected features. Considering these characteristics, we proposed the cluster particle filter to improve the global localization robustness and accuracy. Experiment results show the effectiveness and robustness of our approach in our service robot ApriAlpha™ Platform.