Research On Indoor Localization And Navigation Based On Speech Recognition

An explosive growth of mobile devices is accelerating a demand for Location Based Service(LBS). Localization technology is the key of location-based services. In many indoor and outdoor environments, if you can get the user’s accurate location information,you can provide users with the applications of interest and services. In the outdoor environment, GPS can provide perfect localization services, but in the indoor environment,because of the particularity of the indoor environment, most of the locating and tracking methods rely on constraints of indoor floor plans, which requires a large number of exploration and greatly reduces the scalability. In view of this, this paper fully studies the correlation of pedestrian’s communication content and spcace then proposes an indoor localization and navigation method based on Automatic Speech Recognition(ASR) and received signal strength(RSS) sequence of multiple APs. The main contents of this paper are as follows:Firstly, we include the methods and main principle of indoor localization and navigation based on smart phones, then we summarize the main constraint factors to apply these methods.Secondly, we analyse the correlation of the location semantics and the space, and propose a coarse granularity localization method based on ASR, then draw into the Conditional Random Field(CRF) to estimate the semantics of the user’s current location.Thirdly, we propose a localization method based on the location sensitivity of RSS sequence of multiple APs. Sorting multiple strong RSS values as a fingerprint sequence,and locating accurately by RSS values dislocation matching. In addition, we unite the dead reckoning and particle filter algorithm to realize the navigation.Lastly, we design complete experiment to evaluate our system and prove the feasibility of the methods proposed.