With the growing popularity of electronic devices, Graphical User Interfaces (GUIs) are becoming increasingly important. GUI testing has also been prevailing in software testing. However, existing automated GUI testing tools mostly rely on framework of a specific platform. Testers have to fully understand platform features before developing platform-dependent GUI testing tools. Starting from the perspective of tester’s vision, we observe that GUIs on different platforms share commonalities of widget image and layout design, which can be leveraged to achieve platform-independent testing.

In this paper, we propose UniRLTest, an automated software testing framework, to achieve platform independence testing. UniRLTest utilizes computer vision techniques to capture all the widgets in the screenshot and constructs a widget tree for each page. A set of all executable actions in each tree will be generated accordingly. UniRLTest adopts Deep Q Network (DQN) to the exploration process. We encode each widget tree as a state, each executable interactive action as an action and novelly propose a curiosity-driven reward function to guide the agent to those low exploration pages. An experience replay memory is set to sample randomly and train the neural network. We have conducted evaluation experiments on 25 applications from different platforms. The result shows that UniRLTest outperforms baselines in terms of efficiency and effectiveness. The demo can also be found at https://youtu.be/m6htfJ5xdaU.