在移动互联网时代，移动智能终端在众多领域已通过移动应用提供服务。然而，移动应用的测试通常因为开发者追求快速迭代和发布而不够充分，难以确保移动应用能否提供可信服务。而且，平台依赖特性使得针对多平台移动应用的测试因复用度低而浪费了大量资源。本课题研究模型驱动的移动应用测试方法，面向多移动平台实现测试生成与复用，主要研究内容包括：研究面向测试的移动应用与平台的解耦方法、移动应用平台无关行为模型的自动重构方法；研究模型驱动的移动应用测试方法、基于平台无关测试自动构造平台相关测试的测试转换方法；研究针对移动应用特定缺陷的平台无关测试生成方法与跨平台测试执行方法；研制兼容各类移动应用硬件设备、具有跨主流移动平台能力的模型驱动的移动应用测试支撑工具原型，并开展实例研究。本课题的主要创新点在于面向测试自动重构移动应用的行为模型，使得模型与模型驱动的测试方法能够支持解决多平台移动应用的测试生成与复用问题。

In the era of mobile internet, mobile devices provide service via mobile apps in many application area. However, the testing of mobile apps is always not enough because of the rapid develop and release. In a result, it is difficult to ensure that the mobile apps can provide dependable service. In addition, platform specific testing of mobile apps reuses little and wastes a lot, raises new challenges. The project focuses on model-driven testing of platform-specific mobile apps on multi-platforms. In order to test platform-specific mobile apps, we will come up with decouple approaches for separating mobile apps from its platforms, and reconstruct behavior models by extracting platform-specific models from implementations of mobile apps, and then transforming to platform-independent models. We generate platform-specific tests based on the platform-independent tests based on the models of mobile apps. We will work out defect-specific platform-independent model-driven testing approaches for mobile apps. Based on the above research, we will develop model-driven testing framework for mobile apps, which can support multi-platform and multi-device. The main contributions of the project include test-oriented model reconstruction approaches, and the model-driven testing approaches by improving the reuse of tests for mobile apps on multi-platform.