Multi-modal sensor calibration on board the ATLASCAR2

• Main Author: Castro, Afonso de Jesus Cardoso Pinheiro de
• Publication Date: 2019
• Format: Master thesis
• Language: eng
• Source: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Download full: http://hdl.handle.net/10773/29127
• topic: Extrinsic Calibration; Multi-Modal Sensors; ROS; Optimization; Intelligent Vehicles

Complex robot systems have several sensors with different modalities. In order to estimate the pose of these various multi-modal sensors, some works propose sequential pairwise calibrations, which have some inherent problems. ATLASCAR2 is an intelligent vehicle with several sensors of different modalities. The main goal of this work is to calibrate all sensors on board the ATLASCAR2. A ROS based interactive and semi-automatic approach, that works for any robot system, even the most complex ones, was developed. After the step of identifying which geometric transformations, between all robot description, should be estimated and collecting the detected data from each sensor, a least-squares optimization occurs to enhance the position and orientation of each one of the robot sensors. Results show that the four sensors simultaneous calibration is as good as the pairwise procedures used with the standard calibration tools, such as the OpenCV ones. In that way, the proposed solution brings a novel and advantageous methodology, since it fits to any complex robot system and calibrates all sensors at the same time.