Indoor Location Algorithm Based On Wireless Networks

With the rapid development of wireless network technology, the demands of Location-based Service are growing. However, the research of indoor positioning is the emphasis and difficulty. In recent years, a growing number of researchers are engaged in the study of indoor positioning. Due to the complex indoor environments, has not a good algorithm can be applied to all indoor environments. First, this paper introduces some of the existing classic indoor positioning algorithm and their advantages and disadvantages. Then, proposed some improved algorithms based on these algorithms.The innovations of this paper and the main works are as follows:Response to the problem that antenna are non-omnidirectional, this paper proposed a positioning algorithm based on Unscented Kalman Filter in chapter 3. Use of RSSI values to calculate the distance based on ranging model, and use the Unscented Kalman Filter algorithm to estimates node coordinates. Since the RSSI value measurement and ranging model parameters affected by the environment, and the environment is dynamic changes in real-time. Therefore, this paper uses Gaussian filter to optimize RSSI value, uses linear regression algorithm to optimize the environment parameters and uses adaptive mechanism to update.As wireless signal is affected by reflection of walls and objects, diffraction, multipath and antenna directions, using traditional position algorithms cannot achieve well positioning accuracy. A large number of experiments show the many-many relationship between the RSSI value and the physical distance. Response to this inherent characteristic, this paper proposed an algorithm based on n-tuple distance sets in chapter 4. Establish mapping relationship between RSSI values and distances, use Gaussian filter to optimize RSSI values, by calculating the similarity of RSS between blind point and datasets to obtain the mapping distance sets, and using K-means clustering algorithm to classify distance set to reduce the computational complexity.Due to existing position algorithms didn’t consider the important factor of the antenna measuring angles, this paper presents a feature regional positioning algorithm in chapter 5, discusses the measured angles between partitions, and uses intelligent algorithms to identify feature points in the region according to relationship between data which in the same angle interval. By calculating the Euclidean distance between the positioned points and Aps and feature points, we can determine the region which the point to be positioned,In this paper, we use CC2430 nodes and TP-Link routers as fixed terminal to collect data, use Android smartphones as mobile terminals for collecting reference points’ data. Use the collected data to position through this paper proposed algorithms, positioning accuracy has varying degrees of improvement.