Research On Filtering For Range-free Indoor Localization System

With the development of Mobile Internet and Internet of Things, the special status and importance of indoor localization research has increased remarkably. For the application of indoor environment such as shopping malls, hospitals and residential areas, object's location becomes one of the crucial context information. And accuracy of indoor localization systems determines the quality of application layer service. Researches on indoor localization techniques also boost the development of Mobile Internet and Internet of Things. So there is important academic meaning and application value in the research of indoor localization techniques.Indoor localization techniques can be classified as range-based techniques and range-free techniques. Range-based techniques are more accurate but may not be applied easily due to the requirements of specific devices and line-of-sight features. Range-free techniques usually utilize wireless signal strength as location information which leads to a fluctuation in the accuracy because the signal is vulnerable to the interference. There is supposed to be an improvement in the accuracy of range-free techniques if the environmental information and trajectory of moving target can be take into consideration effectively. For these issues of range-free indoor localization techniques, this paper issues relevant topics and gets following results:1. Based on the idea of multi-source information awareness, this paper defines a modeling method of multi-source information and testifies the model through typical fingerprint matching algorithms which proves the effectiveness of this model and its improvement in location estimation accuracy.2. In consideration of that moving target's trajectory is continuous, this paper studies relevant issues of applying filter techniques to the tracking problem and summarizes the designing scheme of state transition model. The designing scheme of noise model is also optimized through an experimental analysis.3. Based on the above research results, this paper designs and constructs an indoor localization prototype system. This system uses multi-source information model and utilizes Bayesian filtering method. An experiment in real scene is carried out and the results show that this system improves the accuracy comparing with systems without filtering techniques.