Probabilistic Calibration of a Manipulator and a 2D Laser Scanner

An accurate geometric calibration between the subsystems of a mobile platform is essential for a multitude of applications. The laser scanner is of particular relevance for tasks such as localization and mapping, as well as for object- and obstacle detection. Thus, the calibration between a laser scanner and a manipulator is of particular interest for mobile manipulation and relevant for the use in industrial applications, disaster szenarios or search-and-rescue missions. In this paper, we present a probabilistic method for geometric calibration of a manipulator and a 2D laser scanner. In contrast to state-of-the-art procedures, the presented method takes into account the uncertainties of the measurements of the laser scanner and of joints which is propagated through the kinematic chain of the manipulator. Furthermore, our approach provides an estimate of the uncertainty of the resulting calibration. The method does not require any additional sensors and constitutes an accurate and easy-to-apply solution. The approach is validated with data of simulation and real experiments. We show in numerous monte carlo simulations that considering the uncertatinties improves the accuracy by 50.2% in translation and 44.3% in rotation compared to a state-of-the-art procedure. The plausibility of the estimated calibration is validated by real experiments. Furthermore, the provided uncertainty of each calibration parameter indicates whether recalibration is required.