Model based code generation for kinematics and dynamics computations in robot controllers

This paper gives an overview on our current research on models and Domain Specific Languages (DSLs) for robotics software development. We target the set of kinematics and dynamics algorithms which are critical for simulation and control (e.g. forward kinematics or inverse dynamics), but whose manual implementation is error prone and time consuming. Our approach relies on model based code generation to address the apparently conflicting requirements arising from the constraints of real time robot controllers on one hand, and the ease of use and the flexibility desirable by the end user on the other hand. Evidences from our experience and numerical experiments demonstrate the effectiveness of this approach.