Machine Vision Cannulation of Central Veins

Ultrasound will be the mode of vessel localization and dynamic guidance of needle insertion. An acoustic coupling for the ultrasonic sensor is accomplished with alcohol based gel which also serves to decontaminate the insertion site. Current ultrasound technology is capable of delivering near real-time data and thus compensates for environmental motion, shocks, and vibration. Unlike a haptic control system, ultrasound guidance will not be negatively affected by environmental motion artifacts which are inherent in a moving vehicle or in a combat situation.

IJ/FV cannulation is a routine medical procedure, performed on a daily basis in elective medical situations in large numbers. This means that all the elements of the proposed device can be developed, evaluated, and validated in standard medical practice situations.

The ultrasound sensing, the machine vision target acquisition, and the coupler to join ultrasonic sensor and cannulation device ( 5.25” 14ga cannula with needle and guide wire) can be used with manual manipulation in the first phase to allow assessment of the motion axis requirements necessary for full automated implementation.

Specific Aims

Develop machine vision target acquisition software for the internal jugular / femoral veins. Utilize multiple images from standard medical care settings, and develop an ultrasonic effector head that incorporates a standard hand-held transcutaneous probe with sufficient resolution to allow accurate and rapid machine vision targeting. This head would couple to a cartridge containing a standard 14ga 5.25” intravenous cannula with 0.032” guide wire. The effector head will be designed for use both as a hand held device as well as a robotic fixture.

Develop a cartridge to hold the above disposable IJ cannulae.

Deploy a standard portable ultrasound machine coupled with a PC with video analog/digital conversion and running standard machine vision software.

Develop specifications for motion control based on hand held applications of the above devices in large numbers of patients undergoing elective central venous cannulation.

Link via gigabit Ethernet the machine vision computer with the motion controller module and deploy a standard 5 axis robot to test the entire assembly in animal models.

Test the refined robotic assembly in human volunteers.

Further test the assembly in human patients undergoing routine internal jugular cannulation.

Test the further refined assembly in the trauma setting.