Researches On Wireless AP Deployment Optimization On Big Data Of User Experience

With the continuous expansion of wireless WiFi coverage network,and the number of users has increased sharply,the problems of WiFi that represented by colleges and universitiesis particularly prominent.As an important device in a wireless network,wireless AP(known as Wireless Access Point)is closely related to the wireless Internet experience of the users who use campus network to access the Internet.The factors such as,the coverage of wireless AP,arrange of location,transmit power,antenna orientation,and placement distance between APs all affect the wireless Internet experience of campus network users.The sparsity of AP will cause signal strength to attenuate with increasing distance.Meanwhile,AP frequency or power overlap,building occlusion,a large number of user groups gathered in the same AP and other adverse environments will also reduce the users' Internet experience.Because of the characteristics of WiFi,and affected by the busy degree of the channel.It is not that the denser the AP,the better the user experience.Therefore,by optimizing the deployment parameters of wireless APs,it can better meet user needs,thereby improving the wireless Internet access experience of campus network users.However,the current optimization method for wireless AP deployment parameters is mainly based on empirical manual attempts,which cannot be widely applied.Based on the user data generated by wireless Internet access by teachers and students of Northeast Normal University,this thesis introduces big data technology into the process of wireless AP optimization deployment,and designs and realizes the process management system of wireless AP adjustment of campus network.The system uses the Hadoop platform to perform cluster analysis on user experience big data,so that analyze the distribution of signal attenuation and superposition of each wireless AP in space.And,according to the results of big data clustering,combined with the corresponding mathematical model of wireless AP optimization deployment given in this article,gives the deployment adjustment suggestions for a specific AP.There are two main problems in this thesis.First,determine the coverage and signal strength distribution of wireless APs based on user experience big data,and model it as a big data clustering problem.Second,the wireless AP deployment optimization scheme is given based on user experience,which is an optimization problem in engineering design.How to through the wireless AP signal strength,time space and other variables of the constraints,so that the user experience to achieve optimal value,which requires repeated optimization and real measurement.The following three main points of research have been done in this thesis:(1)Design and implement wireless AP adjustment process management system based on Hadoop platform on campus network to provide technical support for campus network center staff;(2)Design and implement a parallel k-Means ++ clustering algorithm based on MapReduce,also build a Hadoop cluster to perform cluster analysis on user data,visually display the signal strength and coverage of the entire campus,so that managers can intuitively understand the wireless Internet experience of teachers and students;(3)Give the data model for optimized deployment of wireless AP,which is beneficial for technicians to adjust the parameters of wireless AP devices in a timely manner,so that improve the wireless Internet experience for college users.