Spherical multi-directional eddy current sensor for robotic profiling of metallic edge surfaces

Accurate profiling of metallic edge surfaces is critical for ensuring structural integrity and high-precision assembly. This study presents a robotic scanning system that incorporates a novel spherical eddy current array sensor with nine spatially distributed coils, offering multi-directional sensitivity for adaptive edge surface profiling. The system utilizes mutual inductance measurements as feedback to perform real-time sensor pose adjustment, maintaining the sensor axis perpendicular to flat regions and enabling continuous transitions across edge geometries. The resulting sensor pose is also used to estimate the edge direction, enabling integrated reconstruction of both geometry and orientation. The sensor design is informed by an analytical model of tilted coils, with the coil arrangement optimized according to the spatial distribution and amplitude of sensitivity. Experiments on copper foil samples with both inner and outer edges at 120°, 90°, and 60° demonstrate accurate reconstruction results. The acquired scanning trajectories closely follow the true surface profiles and exhibit symmetric arc transitions near edges. Experiments under different lift-off thresholds show that increasing lift-off produces a wider transition but increases the maximum deviation from a sharp corner, and that lower material conductivity leads to a broader edge response. For a 90° edge, the estimated edge orientation deviates by less than 1° from the ground truth. • A novel spherical 9-coil eddy current sensor with multi-directional sensitivity is developed for metallic edge inspection. • The sensor’s spatial coil arrangement is optimized based on an analytical model of tilted coil sensitivity. • An adaptive robotic system integrates the sensor to maintain optimal pose during edge surface scanning. • Experimental validation on copper foil edges demonstrates accurate profile reconstruction and edge orientation estimation with errors within 1°.