Indoor Positioning Based On Multi-Sensor Of Smart Phones

With the popularity of smart phones and the development of computer technology,the demand for positioning service is becoming more and more intense.And about the positioning technology,the outdoor positioning technology,GPS,has been able to provide stable and efficient positioning service.With the rapid development of smart home,indoor positioning technology has attracted more and more attention.At present,most of the popular technologies in indoor positioning rely on WiFi,geomagnetism,Bluetooth,infrared and so on.Among them,with the advantages of resources,indoor positioning technologies based on WiFi and geomagnetism have obtained more and more attention,and both of them have their own defects working alone,and using a single sensor,can't fully eliminate the limitations of positioning.Therefore,this thesis uses the intelligent mobile phone to collect the WiFi signal and the magnetic field signal in the room,combines the two to use the particle filter algorithm to locate.Most of existing particle filter based approaches either reduce the effective particles after resampling or are heavily affected by motion estimation errors,which result in unreliable systems.Aiming at the problems existing in particle filter location algorithm,this paper proposes an improved particle filter algorithm through improving the importance sampling process,and the simulation experiment proves that the accuracy of the proposed localization algorithm is improved.The main works of this thesis are as follows: the signal model is described,then the state space model of dynamic system,Bias principle,Monte Carlo method,Bias importance sampling and sequential importance sampling are introduced.The characteristics of the smart phone sensor are introduced,and the influencing factors of the positioning signal are analyzed,and the corresponding solutions in the experiment are described.From the importance sampling process and resampling process,the principle and process of particle filter algorithm are elaborated,and the defects of particle filter algorithm are analyzed in detail.Then the state transfer equation and the prediction decision are designed from the motion model of the target.It also introduces an improved evolutionary strategy in the process of resampling.The design of the system,the preparation of the experiment and the analysis of the results are introduced.