Reasarch Of Smart Home System Indoor Positioning Technology Based On IOT Gateway

With the improvement of people’s living standards and the development of sensor and network communication technology, smart home is gradually into people’s daily life. However, the traditional system of smart home usually just control home furnishing equipment and detect home environment, which is based on embedded technology, sensor technology and integrated wiring technology, etc. In recent years, the detection of human body information such as the number and position in the home environment through indoor positioning technology is increasingly concerned. As an important technology in smart home area, indoor positioning technology is a key way to improve the comfort and safety of the home environment and the degree of intelligence.According to the positioning requirement of indoor human body and objects in smart home, this paper mainly focuses on indoor positioning algorithm based on radio frequency technology. After comparison of some indoor positioning methods based on radio frequency technology, this paper focused on the typical LANDMARC indoor positioning system and its nearest neighbor algorithm, which is based on RSSI methods. For lack of the nearest neighbor algorithm, we proposed a novel algorithm to improve positioning accuracy. The main research achievements of this paper include the following aspects:Firstly, this paper researched on the typical LANDMARC indoor positioning system. After briefly introduction of the system and its energy level calculate method, we focused on analysis of principle of the nearest neighbor algorithm, and simulate the algorithm using MATLAB. Simulation results showed that the number of nearest neighbor reference tag k and the density of reference tag will both influence accuracy of the algorithm. We can come to the conclusion that the nearest neighbor algorithm achieved best performance when k equals4, the maximum positioning error is1.18m, and the average positioning error is0.74m. In addition, properly increasing the density of the reference tag can improve positioning accuracy of the algorithm.According to the short advantage of LANDMARC positioning algorithm, this paper proposed a novel algorithm based on BP neural network, we named it BPNN-LANDMARC. We first introduced principle of the improved algorithm, and then simulated it using MATLAB. Simulation results showed that the proposed algorithm can improve the positioning accuracy. Its maximum positioning error is0.89m, which is decreased by24.57%, and the average positioning error is0.56m, which is also decreased by24.32%.Finally we built a smart home gateway platform based on S3C6410processor and embedded Linux operating system. Then we transplanted the positioning algorithm into the gateway and designed multithread sever and serial port program. Based on the designed hardware and software platform, we tested and verified the positioning performance of the proposed algorithm and the original algorithm. The actual experimental results showed that the proposed algorithm in this paper can improve the positioning accuracy.