The development of Human Robot Collaborative (HRC) systems faces many challenges. First, HRC systems should be adaptable and re-configurable to support fast production changes. However, in the development of HRC applications safety considerations are of paramount importance, as much as classical activities such as task programming and deployment. Hence, the reconfiguration and reprogramming of executing tasks might be necessary also to fulfill the desired safety requirements. Model-based software engineering is a suitable means for agile task programming and reconfiguration. We propose a model-based design-to-deployment toolchain that simplifies the routine of updating or modifying tasks. This toolchain relies on (i) UML profiles for quick model design, (ii) formal verification for exhaustive search for unsafe situations (caused by intended or unintended human behavior) within the model, and (iii) trans-coding tools for automating the development process. The toolchain has been evaluated on a few realistic case studies. In this paper, we show a couple of them to illustrate the applicability of the approach.

人机协作（HRC）系统的开发面临许多挑战。首先，HRC 系统应该具有适应性和可重新配置能力，以支持快速的生产变化。然而，在 HRC 应用程序的开发中，安全考虑与任务编程和部署等经典活动一样重要。因此，执行任务的重新配置和重新编程也可能是必要的，以满足所需的安全要求。基于模型的软件工程是敏捷任务编程和重新配置的合适手段。我们提出了一个基于模型的从设计到部署的工具链，它简化了更新或修改任务的程序。该工具链依赖于 (i) UML 配置文件以进行快速模型设计，(ii) 对模型中不安全情况（由有意或无意的人类行为引起）进行详尽搜索的形式验证，以及 (iii) 用于自动化开发过程的代码转换工具。该工具链已根据一些实际案例研究进行了评估。在本文中，我们展示了其中的几个来说明该方法的适用性。