The Research And Application Of Indoor Positioning Based On Affinity Propagation Clustering Algorithm

At present, with the vigorous development of the mobile internet,the service based on location become more prevalent in people's daily life. It displays its tremendous value in such areas as goods storage,emergency aid, transportation management and personalized information process, which greatly facilitated our daily life.The outdoor positioning system, mainly relying on cellular network and global positioning system, has already been very mature and stable in the practical application of outdoor positioning, but it will stop working on the condition that environment is blocked by indoor buildings. Therefore,indoor positioning technology is badly in need. The indoor positioning system based on WIFI is a main object for research due to its fast deploy and low cost. However, the fluctuated indoor WIFI signal, unavoidable noise and natural barrier mean that there will be a large room for the improvement of the indoor positioning system.By improving the location fingerprint positioning technology, the thesis optimized each stage of the indoor positioning technology, mainly including : pretreatment optimization of signal data during the offline stage; the improvement of positioning accuracy by eliminating abnormal points with the help of the Affinity Propagation. The thesis coves the following achievements:1. By focusing on the accuracy of the location-based fingerprint positioning algorithm, the thesis put forward the improved version based on Propagation Algorithm. Firstly, during the off-line stage, data of collected signal was denoised through filtering process; secondly, during the stage of online positioning, positioning points were clustered by Propagation Algorithm. The biggest cluster is where positioning results lie, and its center is regarded as the final result. The experiment show that the final positioning result is more accurate than currently mainstream positioning algorithms, such as NN, KNN and WKNN.2. The indoor positioning system based on WIFI is designed and realized. The core module in this system is that the server realized accurate positioning by using the positioning algorithm put forward by the thesis. The system concerned mainly includes the Android Client Side and the Server Side, the former is used for data collection and online positioning, and the latter for data storage and calculation of positioning results. On the system from the aspects of functional,robustness and so on were tested, the results show that can satisy the system requirements, to achieve a more accurate positioning.