Research Of Indoor 3D Positioning Technology Based On Improved Neural Network

Indoor environment plays a very important role in people's production and life.With the rapid development of Internet and Internet of things,the coverage of the network is expanding,and the demand for location aware services in the indoor environment is becoming more and more large.The outdoor location service dominated by GPS has been far from meeting the needs of people.With the continuous expansion of the application scene,people have become more and more urgent for the location perception service in the underground parking lot,the supermarket and many other indoor environments,and the particularity and complexity of the indoor environment have also led to the difficulties in the study of indoor three-dimensional positioning technology.How to overcome these problems also becomes one of the hotspots of the present research.In view of the above problems,this dissertation first expounds the research background and significance of Indoor 3D positioning,introduces several classic indoor positioning methods,and then describes the application of traditional methods in indoor positioning.In the light of the research status of universities and scholars at home and abroad in recent years,a scheme of Indoor 3D positioning system with mobile phone as terminal and covering Wi-Fi signal as application scene is proposed.In view of the system,this article mainly completes the following parts:(1)Compare the current technology research with traditional methods,compare their advantages and disadvantages in different scenarios.In view of the complexity and particularity of indoor environment,this dissertation discusses the more popular neural network model and particle swarm optimization in recent years,and analyzes the theory of model and algorithm thought,and studies its feasibility in indoor positioning technology.(2)Combined with the advantages of neural network and particle swarm optimization,an improved multi strategy particle swarm optimization algorithm is proposed to optimize the topology and weight threshold of neural network.The optimal relationship between the Wi-Fi signal receiving intensity RSSI and the receiving distance D is captured by the nonlinear processing capability of the neural network,which not only improves the universality of the algorithm in different environments,but also improves the positioning accuracy.Finally,a comparative analysis with BP neural network and genetic neural network is carried out to prove the feasibility and effectiveness of the method.(3)Combined with the improved 6 point weighted centroid algorithm of space ball,the coordinates of measured nodes are solved.The algorithm adds a launching node to the original basis,which makes the space range of the node to be measured further narrowing,the node positioning accuracy is further improved,and compared with the traditional quadrangle measurement method and the centroid method.(4)Based on the Android platform,the indoor 3D positioning algorithm is designed,and the client and server functions are designed according to the system architecture and localization process.