Research On Indoor Positioning Algorithm Based On Fuzzy Inference

In recent years, with the development of wireless communication technology, people’s demand for application of location services in the navigation, first aid and information transfer areas has increased rapidly. Companies and experts from both domestic and overseas have done countless researches in the field of positioning and already made remarkable achievements. In the outdoor environment, the GPS(Global Positioning System) has already been very mature to meet most of the positioning and navigation requirements.However, for the case of indoor environment, it does not work well. Firstly, the satellite signal can’t penetrate buildings; secondly, the complicated indoor environment causes serious multipath effect and interference of non-line-of-sight. These two reasons make it hard to use outdoor positioning method to get a satisfying result of localization. Therefore, it is necessary to use some specific algorithms to solve the indoor positioning problem. Because of the difficulty of establishing a high precision signal propagation model in the indoor environment, the parameterized method is unlikely to solve the problem. So the usage of non-parametric method is chosen correspondingly, for instance, the fingerprint positioning method. Due to the indoor environment signal’s heavy irregularity and uncertainty, this thesis used fuzzy theory and then realized the indoor positioning algorithm based on fuzzy inference.The main content and principal contribution of this thesis are as follows:(1) By researching related documents and studying indoor positioning technology, this thesis analyzed the advantages and disadvantages of the existing indoor positioning technology, and determined the research object is fingerprint positioning algorithm based on received signal strength.(2) Some aspects of fuzzy theory has been introduced and been combined with the indoor positioning algorithm. And how these provided theoretical support of the implement of the new indoor positioning algorithm based on fuzzy inference.(3) Made the realization of using membership function of fingerprint to train the fuzzy inference system and obtained the confidence level of fuzzy rules. By building fuzzy inference system and using the Newton’s method to solve a set of nonlinear fuzzy equations, this thesis completed the design and implementation of the new indoor positioning algorithm based on fuzzy inference. Then the principle of the algorithm and implementation steps was described in details. And simulations in different scenarios were done to prove the validity of this algorithm.(4) A brief summary of the whole paper, and put forward the next optimization and improvement of further research.