A Teleoperative Ultrasound Robot System with Low-Cost Method for Real-World Image Acquisition

Ultrasound is widely used to diagnose human organs because of its low cost, no harm and high imaging efficiency. What’s more, there are some situations that human interaction can barely reach, like COVID and radiotherapy. Previous studies utilized Inertial Measurement Unit (IMU) or 6D haptic device to transmit posture data to conduct the robotic arm scanning. However, there exists limited portability for it highly dependent on specific devices.

Here we are committed to developing a semi-automatic robot system for ultrasound scanning, which proposes a point cloud rotate based method to locate the organs in multi-view. In addition, we first introduce 6D object pose estimation into robotic ultrasound to facilitate the portability of human-robot interaction. By manipulating a carry-on object and synchronizing its 6D pose to a robotic arm, one can get rid of the constraints of a specific device. To ensure patients’ safety and the image quality during scanning, we apply admittance control, which keeps the probe constant contact with human issues and adaptive to patients’ motions. The system first locates the organs in a rotated point cloud image, then attaches to patient’s body automatically with a certain force and finally finishes the scan with the help of a remote doctor holding a portable object. Experiments are conducted to validate the proposed method.