Abstract:In view of the fact that most floor identification methods currently do not consider the activity recognition of up and down stairs resulting in the floor identification results switch back and forth in the stairwell.This paper proposes a floor identification method based on Wi-Fi/barometer combination. In the offline phase the fingerprint database is established on the plane floor, and the fingerprint database is clustered by the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm to distinguish the signal characteristics of different regions. In the online phase the receiving signal and the fingerprint database are matched to identify floor. In the stairwell, the barometer is used to identify the activities of the upper and lower stairs. The activity recognition incorporates the step frequency detection to cope with the misjudgment caused by the sudden change of the motion state. The results show that the accuracy of the method of this paper is obviously improved compared with the basic floor recognition method in the plane floor. The floor transition can effectively identify the up and down stairs activities, and solve the problem of the switching between floors in the stairwell. After the step frequency detection is incorporated,the misjudgment caused by the sudden change of the motion state can be effectively eliminated.Experiments show that the floor identification method of this paper can effectively deal with the floor identification requirements of complex environments, and the completeness is strong.