Component-based modelling for generating robotic arm applications running under OROCOS middleware

Nowadays, complex dynamic systems, such as robotics, demand new and highly sophisticated software developments. But, basically, robotics arm applications need to get information, through sensors, from the variant and unknown environment; this is available to controllers which also have information about the robot's state (position, velocity, acceleration, torque, etcetera) and taking into account a reference, they generate the next state for the robot. Therefore, every robotic platform should require a set of sensors to take information from the environment, controllers and robots (with the low level controller). In order to cope with devices, which part of them exhibit vendor-dependent characteristics, a standard interface is proposed to allow the designers to abstract the vendor specific drivers' knowledge and to guide them in this work. This paper also explores the advantages of model-based techniques to design robotic-arm applications. In particular the Unified Modelling Language has been used for modelling the functionality in the robotic's manipulator applications, from where components of the OROCOS (Open RObot COntrol Software) are automatically generated. The proposition, characterized by its openness, productivity, portability and interoperability, is based on the construction of a complex robotic manipulator platform that integrates a Staubli industrial robot and a MEKA anthropomorphic robotic arm. Both can be used, in an easy way, in advanced experiments as it is shown in the results.