Design And Implementation Of Android Application Security Detection System

With the increase of the market share occupied by the Android system, security issues become the fetters of the Android system for further development. Although the Android system provides a sandbox mechanism, permission mechanisms, and digital signature mechanism, Google didn’t provide strict strategy to manage the application market and filter out malware which seriously affected systems and safety of users at the entrance. Therefore need for a new application security detection mechanisms to detect the applications to ensure the reliability of the applications.Analyze the Security of the Android application by the application’s runtime behavior. By researching the system security mechanisms and malware, abstract malware’s pivotal behavior and associate application’s information form application pivotal behavior sequence. By training a BP neural network to evaluate the behavior of the application, this evaluation method is not only able to identify known application behaviors of malware or unknown malware also has a good effect.The main assignment of this thesis is to design and implement an Android application detection system. The overall structure of the paper is divided into seven chapters. Chapter I summarizes the security problems faced by the Android system and the reasons and put forward the overall structure and the research contents; Chapter II analyzes the security mechanisms provided by the Android system and security threats, and summarize the malware through intrusion, dissemination and intent; Chapter III, In-depth researching permission management mechanisms and associate permissions with application’s behavior, By dynamic detection mechanism and decompiled code injection technology to achieve the capture of application’s behavior; Chapter IV, given the overall design of the system and presents an assessment of the behavior of BP neural network model based on the definition of the extended application of behavioral sequences by known malware software applications and software applications normal behavior information to train the BP neural network; Chapter V, the application for Android detection system for a detailed design and implementation; Chapter VI test, the system was functional and performance testing to prove the availability of the system; In the end, summarizes the paper work and the outlook for further work.