Combining control design tools - from modeling to implementation

Good tools are needed in order to develop robotic systems efficiently. Today, in addition to CAD/CAM, there are tools for model derivation, control design and implementation. There are also tools for exporting models to a control design environment, as well as from control design to implementation (rapid prototyping tools). It is however, still difficult to combine these tools, especially when working with large systems (i.e. a lot of signals and parameters). We have therefore combined, interfaced and augmented some of these tools into a method that bridges the gaps between automatic model derivation and control implementation. Analytically derived functions (from Maple) are used for control design, simulation, visualization, evaluation (MATLAB) and implementation (Real-Time Workshop and xPC Target). The method and tools are illustrated by application to a four-legged robot in both simulation and reality. Generality is also examplified by applying the method to a simulated SCARA robot.