Tissue-mimicking Torso Phantom


Project status: Active


Despite many advancements, ultrasounds remain a primarily manual process. Robotic ultrasounds offer patients better precision, consistency, and efficiency. They also reduce the risk of musculoskeletal conditions and fatigue for sonographers who would otherwise be doing back to back ultrasounds for prolonged periods of time with non-ergonomic transducer designs. Unfortunately, full automation of even basic ultrasounds still faces challenges. Due to the inherent squishiness and inhomogeneous tissues, it’s hard to estimate the position of robotic arms in real time and accurately follow paths. 

Consequently, the development of an autonomous ultrasound system requires a durable and accurate tissue-mimicking model. My goal is to create a silicone model that will outlast the standard of gelatin, while maintaining similar mechanical properties to soft tissues. The tissue phantom will be a torso, which offers unique challenges from the interface of ribs and soft tissues. This area contains many vital organs and is the most commonly imaged with ultrasounds. The physical model will be used to test and measure safety and precision through force and deformation measurements gathered using a UR5 robotic arm’s inbuilt sensors and externally mounted RGBD cameras. The phantom will include a model of the bones, kidney, and a heart to analyze the ability of a robot to accurately produce these images. This was chosen as astronauts face increased risk of kidney stones and heart issues and are in an environment where autonomous medical care is ideal.



Gabriela Morales, Main Engineer and Researcher, Pratt, 2023


Archived Grant Project