Autonomous eFAST ultrasound scanning by a robotic manipulator using learning from demonstrations

Abstract

ABSTRACT² We propose a learning-based controller to enable autonomous execution of the eFAST scanning by a lightweight robotic manipulator according to expert demonstrations. The benefits of this approach are two-fold. Firstly, the automatically acquired USS images can be sent to the expert radiologist from a remote location without the need for complex robotic teleoperation. Secondly, the application of learning by demonstration alleviates the complexity of robotic programming and allows extracting operator-specific knowledge IN SITU in a natural and intuitive way. The provision of incorporating force information can further improve the versatility of the system, allowing easy adaptation to different dynamic environments.

Surgical robotics beyond enhanced dexterity instrumentation: a survey of machine learning techniques and their role in intelligent and autonomous surgical actions

“…In recent papers [90,91,92], GMM algorithms are used to learn from demonstration by representing datasets stochastically using joint probability densities. Kassahun et al [90] developed a method to learn the model of the interaction between catheter and aorta.…”

Semi-autonomous navigation for robot assisted tele-echography using generalized shape models and co-registered RGB-D cameras

“…Therefore, if the target shape is very different from the training set (e.g., people in different poses, very young subjects, people with disabilities), the shape model may fail to register correctly. To address this issue, we can introduce a dividable human model which represents the body as a set of articulated body parts, such as in [9]. For each body part, we can train a deformable shape model based on the statistical shape model.…”

“…Ultrasound examination routines, like FAST, can be preprogrammed on the generic model and automatically adapted to any human body in real time. Additionally, based on a generic body model, sonographerspecific skills and examination habits could be encoded [9], and then autonomously reproduced in the patient specific domain. To the authors knowledge, this is the first study that attempts to address the navigational aspects pertinent to robot assisted tele-echography.…”

Tele-healthcare with humanoid robots: A user study on the evaluation of force feedback effects

“…Besides a collaborative teleoperation mode a semi-automatic mode was presented in which the robot drives autonomously to a chosen location maintaining a constant interaction force with the body. In a similar setup with the LWR the authors of [24] implemented an autonomous system for ultrasonic examination controlled via a learning from demonstration approach. Even the prototype robotic system in [25] for minimal invasive surgery is based on this technology.…”