一种基于APP无缝集成的应用流自动协同框架

An Automatic Collaboration Framework for Integrating Mobile Apps into Flow Seamlessly

应用市场中存在着海量的移动应用(mobile Application,App),无论是在个人日常生活还是在专业领域都发挥着非常重要的作用.每个移动应用都是由不同的提供者独立开发的,提供特定的服务,在设计之初并没有考虑与其他移动应用之间的协同,这种彼此隔绝使其呈现出碎片化特征.面对日渐复杂的用户需求,这种碎片化特征导致用户在多个移动应用间频繁切换才能完成特定任务,用户体验较差.本文重点关注异构自治移动应用间以及移动应用与用户间的协同交互,提出一个基于Android的移动应用自动协同框架,允许在相关的移动应用之间信息共享与复用,并按照应用流的方式逐一执行,从而实现无缝连接满足复杂用户需求.该框架的关键技术包括三个部分:(1)将移动应用的内部数据表示为〈存储,语义,语法〉三元组,并利用静态分析和动态分析技术生成移动应用抽象服务化模型,为异构应用间数据复用奠定基础;(2)使用Nowcasting和Forecasting结合的移动应用预测方法及客户定制方法来构建应用流(mobile Application Flow,App Flow);(3)基于参数注入和Android进程间通信机制设计实现了移动应用间自动协同按需跳转执行引擎.在Android 7.1系统上实现了框架原型,并针对旅行外出典型应用场景展示了异构自治移动应用自动协同的效果.

There are a large number of mobile applications(App for short)in the application market,which play a very important role in the daily life of individuals and in the professional application domains.Each App is independently designed and developed by a different provider,providing a specific service that was not designed to work with other Apps at the beginning.This isolation phenomenon from each other heterogeneous Apps leads to a fragmented feature of Apps.Faced with increasing the complex requirements of users,this fragmentation feature drives users to switch frequently and manually between multiple Apps to complete a specific task,resulting in apoor user experience.In this paper,we focus on the collaborative interaction between heterogeneous autonomous Apps and between Apps and users.In order to meet the complex requirements of users,it proposes an Android-based Apps Collaborative Framework to achieve the seamless connection of Apps,which allows information sharing between related Apps and executes them one by one according to the application flow automatically.The key technology of the framework consists of three parts:(1)Representing the internal data of the heterogeneous App as a set of〈storage,semantics,grammar〉triple,and using the Data-driven static program analysis and dynamic program analysis techniques to generate the abstract service model of Apps;(2)Using the App prediction method combined with Nowcasting and Forecasting and user customized profile(UCP)to build a mobile application flow(App Flow);(3)Designing the automatic collaborative execution engine based on parameter injection and the inter-process communication mechanism in Android to realize the on-demand switch between Apps.The advantages of this framework are mainly reflected in two aspects:(1)Mobile application prediction and recommendation based on application flow avoids the shortcomings of user-defined process,and human-computer interaction based on feedback and inductive learning further improves the accuracy of recommendation results;(2)Mobile application flow execution based on UCP avoids the complex calculation required for application prediction while preserving the flexibility and personalized features of mobile computing,and the sharing of the application process configuration profile among different users can overcome the cold start issue caused by missing user history to some extent.The framework prototype is built on the Android 7.1 system.Based on the prototype,the typical application scenario of travel out is used as a case study to demonstrate the automatic collaboration of the heterogeneous autonomous Apps.The results show that our App Collaborative Framework is effective and promising.In short,the framework proposed and constructed in this paper can effectively break the information barrier between different mobile applications,and greatly improve the user experience in the face of increasingly complexity of user requirements.In the future,we will focus on the following works:(1)The abstract model needs to be further enriched,distinguishing front-end UI,back-end cloud services,and third-party cloud services to be compatible with more interaction requirements and complex business processes;(2)The potential security risks in the implementation process flow will be analyzed and optimized to improve the security and robustness of application flow collaboration.