Research On Wireless Terminal User Identification Algorithm Based On Vision And Wi-Fi Network

In recent years, the combination of wireless positioning technology and video surveillance technology has become a hot issue. In order to solve the problem of target detection and recognition with a small amount of available data, a user identification system based on the moving trajectory matching is proposed in this paper.Existing users discovery and identify methods such as face recognition or electronic tags, own some defects including high cost, limited by the recognition ability and scope, custom hardware and others. Therefore, combining the computer vision and the wireless technology make the new users discovery and identify methods have the advantages of both the electronic tag and the computer vision.In this paper, we locate the users and its carrying wireless terminal respectively, and obtain their moving trajectory and the signal feature of the wireless terminal. In order to identify the user, the algorithm match the wireless terminal with user’s moving trajectory using the signal feature. From the following aspects to solve the problem of user identification. The algorithm solve the identification problem with the following aspectsFirstly, two wireless terminal location algorithms are proposed. We use the smart phone as the user’s electronic tags. Wi-Fi network is used to discover the wireless terminal, and the received signal strength indication (RSSI) is analyzed to complete the positioning and tracking of the wireless terminal. The algorithms take into account the volatility of the RSSI, using the continuous positioning to obtain the signal feature, and match the wireless terminal with the default path which is training in advance. The results show that the matching success rate of the wireless terminal reaches 71%, and the algorithms solve the practical problems such as the difference of the equipment and the electromagnetic environment.Secondly, a matching algorithm based on the moving trajectory of user and signal feature of the wireless terminal is proposed. According to the situation of the test site, a color-based particle filter target tracking algorithm is realized to obtain the user’s moving trajectory. Using the same path used in wireless terminal positioning, the algorithm finish the unification of data format. With the condition of time match, the algorithm complete the binding of the user and its wireless terminal. We implement a minimum prototype system to verify the performance of the matching algorithm, the results show that, the system can reach the user identification rate at 65%.Finally, we use the machine learning algorithm to train our model, which not only gets rid of the limitation of the default path, but also increases the user recognition rate of the system. After using the support vector machine and random forest of two classifiers, the final performance analysis shows that the recognition rate of the system is increased to 87%. Completely satisfy the needs of the users to identify the offline state.