Model-oriented control software development of academic autonomous test vehicle

Abstract

This paper presents a model-oriented toolset developed for academic autonomous vehicle projects. Typical parameters are modeled into a structured domain model, which provides an organized view of the target vehicle. This model can be used to generate the bridge software linking low-level control software (e.g. CAN network) with high-level middleware frameworks. Other software items can be also originated from this model, including the configuration of deployed sensory components and simulation description. Code generation is performed via transforming the vehicle domain model instance into other specific domains (e.g. general kinematic description from vehicle description). As a result, a tool is provided which can be used to define a vehicle configuration in a textual format. The generators create code interfacing the open-source Robot Operating System (ROS).