An Admittance-Controlled Robotic Assistant for Semi-Autonomous Breast Ultrasound Scanning

Abstract

Ultrasound imaging has been shown to successfully diagnose and provide visual assistance during treatment of breast cancer. However, human operators (i.e., technicians and clinicians) provide limited repeatability when performing the imaging scans. Current state-of-the-art automated breast volume scanners (ABVS) have high repeatability but deform the breast tissue significantly, which is undesirable for percutaneous therapies such as brachytherapy. A semi-autonomous system is presented here which leverages the accuracy of a serial manipulator-design robotic assistant to maintain the ultrasound probe at an optimal angle and ensure stable contact with minimal tissue deformation. Positioning of the probe across the surface of the breast is left in the hands of the human operator and is enabled through an admittance controller for the robot. A feasibility study is performed through a comparison of imaging quality for ultrasound scans of a simulated seroma in a phantom tissue when performed with human-in-the-loop and fully-autonomous modalities. The system was evaluated in a user trial showing similar image quality performance to a fully-autonomous position-controlled scanning device.{B}Robot Base Frame{C}3D Scanner Frame{S}Moving Surface FrameB$\vec{P}$Position Vector in Robot Base Frame$c_{\vec{P}}$Position Vector in 3D Scanner Frame$s_{\vec{P}}$Position Vector in Moving Surface FrameB$\vec{V}$Velocity Vector in Robot Base Frame$_ { S } ^ { B } T$Transformation from Surface Frame to Base Frame$_ { B } ^ { S } T$Transformation from Base Frame to Surface Frame