Geometric error modeling and compensation for a tracked autonomous rebar binding robot

Purpose Rebar intersection binding is essential in prefabricated concrete production. Manual binding is labor-intensive and inefficient, while existing rebar binding robots still face insufficient motion adaptability and inherent size constraints in large outdoor rebar mesh environments. This study aims to develop a tracked autonomous rebar binding robot for large-scale rebar meshes and to improve the end-effector positioning accuracy of its robotic arm through geometric error modeling and calibration. Design/methodology/approach A tracked autonomous rebar binding robot integrating a six-track, four-flipper mobile platform, a six-degree-of-freedom robotic arm and an electric binding device is developed. Geometric parameter errors caused by manufacturing and assembly are modeled using a modified Denavit–Hartenberg formulation. An error mapping matrix is derived via differential kinematics. End-effector pose data are measured using a laser tracker, and the geometric parameters are identified and compensated through a least-squares method combined with singular value decomposition. Calibration and on-site rebar binding experiments are conducted for validation. Findings After geometric error compensation, the maximum absolute positioning error, mean error and root mean square error of the robotic arm end-effector are reduced by 70.77%, 70.95% and 69.78%, respectively. Field experiments confirm that the calibrated robot can stably and accurately perform automated rebar binding tasks in real construction environments. Originality/value This work presents a tracked flipper-based rebar binding robot and a complete geometric calibration framework for a self-developed robotic arm, significantly enhancing positioning accuracy and operational reliability in large-scale outdoor construction scenarios.