| With the continuous improvement of pervasive computing research, wireless sensor networks and internet of things, the location service as a kind of new services come into the people's vision. Early location service system is mainly aimed at the outdoor environment. In recent years, with the gradual development of the Smarter City, the Smarter City began to emerge, which the application of indoor positioning has been involved in various fields. However, existing positioning techniques cannot be applied to a variety of complex indoor environment, it mainly reflected in the accuracy is not high, high cost, complex system, adaptive ability is not strong. In view of the above problems, this paper proposed a RFID(radio frequency identification) based devices using the received signal strength indicator RSSI(received signal strength Indication), combined with fuzzy theory and the automatic calibration technology in the actual indoor environment begin to research. Simulation results show that can be applied to practical application.In this work, the RSSI of parametric model in large-scale indoor positioning technology is investigated and studied deeply based on triangulation method. Firstly, the RSSI of distribution characteristics in different scenarios have been analyzed, the data of RSSI processing method combined Gauss filtering and recursive average filtering have been put forward based on the distribution characteristics. And then, in large-scale indoor positioning, the indoor environment can be made region division using the distribution of readers, and fuzzy theory have been proposed to make the coarse positioning quickly based on triangulation method. At last, the three main parameters of the precision positioning model are analyzed, an automatic calibration method have been put forward, the fine positioning automatic parameter model, it are capable to process the real-time parameter calibration in different time and scenes, so as to make the high precision and high efficiency positioning. The major contribution of this dissertation as follows:The distribution characteristics of indoor RSSI are investigated. To describe the RSSI distribution better, the error of RFID based readers and the sample size are analyzed, and the estimated distance between the real distance are contrasted, it provides the experimental basis for enhanced positioning system performance and the method of data processing. The test results show that: the most intuitiveexpression is to reduce the number of 70% samples in the actual data processing, and reduce the computation cost of more than 60% through the distribution characteristics analysis of the RSSI.The distance loss model in wireless communications is studied, and a novel RSSI processing scheme combined the Gauss filter and recursive average filter is proposed. Experiments indicate that compared with the traditional positioning technology in the data processing scheme, this filtering method can greatly reduce estimation distance error and improve the positioning accuracy. In the actual distance error estimation, error at 2m is reduced to 4.2%, and at 3m is reduced to 2.59%.The difficult problem of large-scale indoor location scene is studied, because the RSSI data collected in the real location environment is very large, and the precise positioning is more difficult in one step. So this paper puts forward a coarse location by trilateration positioning method at first. On the one hand, the indoor environment is divided into regions by the distribution characteristics of readers. On the other hand, fuzzy theory is proposed to achieve coarse positioning by adopting the fuzzy distance between the target and readers, it can be concluded as very close, close, acceptable, far, far away. In order to determine the most optimal three readers to realize the coarse location. Test results show that: any indoor environment can realize the coarse location, coarse positioning accuracy is less than 95% at 10 m, and the accuracy are about 23% in the 2m.In order to realize the high precision positioning in different time and different scenes, a parametric model of the large-scale indoor location scene is studied. It mainly analyses the moving individual people in the actual large-scale indoor environment, such as power difference, distance loss factor, and wall influence factor. The localization algorithm of automatic calibration parameters is proposed. Experimental results show that this parameter automatic calibration algorithm can realize the location in 100% compared to the traditional location algorithm, and the positioning error is less than 50 cm, 70% of positioning error is less than 25 cm. More than 80% of the error range is within 30 cm.In conclusion, the indoor positioning system designed in this paper has a certain universality and research significance, it can adapt to the future development. |
PDF Full Text Request