Research On Indoor Localization Technology Based On Wi-Fi Propagation Model

Nowadays, the mobile smart devices are increasingly popular with rapid developing Internet and communication technology. Indoor location-based services has get into everyone's life deeply and has brought great convenience to people's lives. Indoor localization technology has great values in the areas of public safety, health care, location-based services and so on. The improvement of localization accuracy and model adaptability to the environment has become a key problem of indoor localization, which is the current technical difficulties in this field. While GPS localization is the leading technology in outdoor localization, a variety of indoor localization technology presents flourishing state, which is the hot research topic today. Am ong various indoor localization techniques, indoor localization technology based on Wi-Fi has become the most viable ubiquitous localization technology because of its wide coverage, simple deployment, and low cost, and it can fully take advantage of curren t existing wireless access points in the buildings.Existing Wi-Fi indoor localization technology can be divided into these two categories: fingerprint localization and propagation model-based localization. While the former has higher localization accuracy than the latter, but it needs to take significant labor costs for fingerprint collection. The latter reduces labor costs, and it still has the room level localization accuracy but its environmental adaptability is not strong.For the limitations of positioning methods based on Wi-Fi propagation model, this paper studies on localization accuracy and model adaptability, and obtains the following results.First of all, for the case that Wi-Fi signal is interfered obviously by multipath effect in the complex indoor environment, we analyze and study Wi-Fi signal channel state information(CSI), which is the physical layer information of the signal. We can get more signal features for localization through Wi-Fi signal CSI, while this fine-grained feature also helps to eliminate the interference by indoor localization multipath effect, to fit the localization model in the case of overcoming the influence of indoor multipath effect as much as possible.Furthermore, in order to overcome the limitation of environmenta l adaptability of localization method based on Wi-Fi received signal strength indication(RSSI) propagation model, we propose a context-adaptive segmentation heterogeneous RSSI fitting localization method. For the traditional single curve fitting method, we improve it by the idea of segmentation. The experiments in station scene, indoor open scene, and corridor scene these three typical indoor scenes show that our method can make the RSSI distance estimation model more accurate.Finally, for the pros and cons of CSI and RSSI, with the advantage of Wi-Fi signal information at different levels, we design a localization method by fusion of CSI distance localization and RSSI correction. The metho d utilizes the CSI information for localization, and exploits the R SSI propagation model for correction.