Abstract: Aiming to human localization method for elder people assisting in smart home， a method was proposed to realize the indoor localization by pyroelectric infrared sensor (PIR) combined with human behavior-based map. PIRs were used for collecting body motion signal to localize the subject’s raw position in the room. One inertial measurement unit (IMU) was used to collect 3D motion data for human daily body activity recognition such as lying， sitting， and standing. The other IMU was used for velocity and direction estimation. The related behavior-based maps were built by human habits observation. Particle filters method was presented to fuse the data from PIR and IMU，and took into account these prior behavior-based maps to get more accurate localization. Experimental results showed that the proposed approach has the advantage of protecting the privacy while maintaining high accuracy and stability of indoor localization.

Key words: indoor localization, behavior-based map, pyroelectric infrared sensor(PIR), inertial measurement unit(IMU), particle filters