Fig. 7: E-Antenna (larger centimeter-scale) enabled conformal painting for robot arms on curved surfaces.

A Conformal painting experiment setup: (i) A modified brush was mounted on a robot arm to paint on a curved surface consisting of three concaves and three convexes; (ii) An E-Antenna (3 mm in diameter and 24 mm in length) was embedded into the bristles of a paintbrush to synchronously deform with it and provide tactile sensing feedback; (iii) Result of applying 3 N brushing force (labeled as red line) and 3.7 N brushing force (labeled as blue line). B Key frames of robot painting experiments guided by E-Antenna, where \(\gamma\) is the deflection angle of the brush and \({P}_{B}\) is the position of the brush. C The conformal paint result. D Evaluation of painting coverage in these two experiments (blue represents the control experimental results without E-Antenna’s feedback and red represents the expimrntal results with E-Antenna’s feedback), where the E-Antenna enabled conformal manipulation achieves a painting coverage of 99.92%. E The brush trajectories of these two comparison experiments and the surface ground truth profile (black line) recognized in image space, \({\gamma }_{c}\) is the contact angle, d represents the brush’s distance to the flat surface, \({h}_{1}\), \({h}_{2}\) and \({h}_{3}\) represent the height of three convex part. F The deflection angle of the brush. In the experimental group, the brush is controlled to adaptively adjust its posture \(\gamma\) and position to keep the contact angle \({\gamma }_{c}\) constant with the feedback of tactile information, while the brush only moved forward in the control group without E-Antenna feedback. G The brushing force record in these two experiments. In the experimental group, the E-Antenna offered the robot tactile information to guide it to apply constant brushing force on the curved surface, while the brushing force fluctuated obviously in the control group without the E-Antenna feedback.