Safe robot-assisted ultrasound imaging requires a reliable controller able to detect and localize probe--tissue interaction. In this paper, we present a B-mode ultrasound image-based contact perception method and a contact-aware impedance controller for robotic ultrasound imaging. The proposed method detects acoustic contact independently of force measurements, enabling contact-conditioned force/torque taring to reduce residual wrench bias. During contact, the method continuously estimates the effective contact location along the curved probe surface and uses it to update the controller interaction frame, enabling visual servoing of the physical probe--tissue contact point during imaging. Experiments on an agar phantom demonstrated a contact-localization RMSE of 1.46±0.14~mm over probe roll angles from −15∘ to 15∘. During static rolling, the proposed controller maintained task-space tracking accuracy comparable to a conventional fixed-frame impedance controller while reducing the maximum compressive interaction force from 31.56~N to 20.09~N, corresponding to a 36.3% reduction. These results demonstrate the potential of ultrasound images as direct contact feedback for safe and accurate robot-assisted ultrasound imaging.
Figures & tables
Fig. 1 : The experimental setup for RUS system.
Fig. 2 : Contact Point Geometry. A) The probe radius and fixed control frame. The contact frame during B) the non-contact phase and C) the in-contact phase. The red curve is the US probe contact region.
Fig. 3 : US-based Contact Detection and Localization. A) The RUS probe and agar gel. Example US image B) during the non-contact phase. C) when contact is detected, and D) the in-contact phase. E) The contact detection method. The gray shaded area shows the US imaging region.
Fig. 4 : Closed-loop architecture of the contact-aware impedance controller
Fig. 5 : Contact-point Y -position estimation error across probe roll angles.
Fig. 6 : Static controller performance comparison. A) The baseline impedance controller regulates the fixed end-effector tip point, producing additional penetration during rolling (red curves). B) The proposed controller updates the contact point along the probe curvature (green curve). C) Desired and measured probe roll trajectories. D) Independently measured normal interaction forces.
Fig. 7 : Contact-point tracking during linear scanning at different probe roll orientations. (A) Y -directed scanning configuration. (B) Desired and measured contact-point Y position and probe roll angle. (C) Corresponding orthogonal X - and Z -axis tracking errors. (D) X -directed scanning configuration. (E) Desired and measured contact-point X position and probe roll angle. (F) Corresponding orthogonal Y - and Z -axis tracking errors. Shaded regions indicate the roll-transition intervals during which the translational reference was held constant.
Stanford Robotics Laboratory, Computer Science Department, Stanford University, Stanford, CA 94305 USA · Department of Radiology, School of Medicine, Stanford University, Stanford, CA 94305 USA