Abstract

Robotics has adopted modeling with architecture description languages (ADLs). This introduces a gap when reusing solutions encoded in middleware modules. Existing ADL modeling in robotics focuses on domain challenges instead of tool modularity, hence customizing an ADL tool to generate solutions conforming to a specific middleware (e.g., ROS) is challenging. This could produce a multitude of incompatible 'vendor-locked' tool chains and hamper reuse in robotics software engineering. We propose a modular architecture modeling method that rests on the separation of model processing, model transformation, and code generation. This facilitates translating architecture models into modules compatible to the middleware of choice. We present this method using the extensible tool chain of MontiArcAutomaton, which enables translating software architecture models gradually into middleware modules using exchangeable model-to-model and model-to-text transformations. Employing architecture modeling with modular tool chains enables combining the benefits of ADLs with the solutions encoded in popular middlewares and ultimately facilitates robotics software engineering.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call