Abstract: Mobile robot applications often need to self-adaptively deal with the uncertainties in the environments. However, the development of self-adaptive application software for mobile robots is difficult, especially when the application needs to work on different hardware/software platforms. This paper proposes a software framework for self-adaptive robot application systems. The framework reduces the difficulty of the development of adaptive systems, through masking the differences of platforms in hardware and software, and separating adaptive logic from business logic with a rule-based approach. This paper also implements a visualized editor to facilitate the preparation and inspection of rules. The editor can generate model diagrams and automatically check for syntax errors and some logic errors. Finally, this paper evaluates the benefits of the framework with experiments.

Key words: cross-framework, self-adaptive software, rule model

摘要： 自适应软件在移动机器人上具有广泛的应用前景，然而要在具有不同软硬件平台的移动机器人上开发自适应软件是一件非常困难的事情。为了解决这个问题，提出并实现了一个面向移动机器人应用的跨平台自适应软件框架。该框架屏蔽了运行平台的软硬件差异性，并使用规则文件将自适应逻辑分离出来，从而降低了自适应软件开发的难度。为了便于编写及检查规则文件，还实现了一个可视化的规则文件编辑器，用于自动生成模型图，并检查规则中存在的语法及逻辑错误。最后，通过实验测试了框架的使用情况，并对框架的应用效果进行了评估。

关键词: 跨平台, 自适应软件, 规则模型