IEEE Robotics & Automation Magazine - March 2021 - 103

validation in both calibrated mechanical test lungs and live
animals has demonstrated that the VOV is capable of delivering consistent, repeatable, and reliable respiratory therapy
under variable loading conditions.
VOV Development Process
The development timeline of the VOV is presented in Figure 2.
Rallying Cry and Rapid Prototype Iteration
The project began in earnest on 21 March 2020, when physicians at Vanderbilt University Medical Center deemed the
risk of severe local ventilator shortages high enough to make
all efforts that could be brought to bear on the problem. Sensing the urgency in their clinical colleagues, the engineering
team came together quickly-a team consisting of faculty and
graduate students with all of the skill sets required to quickly
build a mechanical ventilator prototype. Within a matter of
hours after this clinical call to action, a napkin sketch made by
one of the engineers [Figure 2(a)] was converted into a first
prototype [Figure 2(b)] that demonstrated the concept of
using a motor-driven mechanism to compress an Ambu bag
at a consistent rate to deliver mechanical ventilation. The
need for accurate, continuous TV adjustment led to the development of version 2.0 on 24 March 2020 [Figure 2(c)]. Version 2.0 implemented the SYM that would become the
preferred transmission mechanism of the design (described
in more detail in the " Mechanical Design " section). In version
2.0, the TV is adjusted by physically sliding the SYM to
increase or reduce the compression of the Ambu bag on a
single stroke. This TV-adjustment mechanism was further

Quick-Release Handle

improved with a manually actuated leadscrew in version 3.0
[Figure 2(d)].
At the time, the system was powered by an off-board,
adjustable lab power supply-meaning that the BPM could be
only crudely adjusted by changing the voltage setting of the
power supply. Realizing the need for more accurate control,
sensors, and safety features, an embedded system (centered
around an Arduino Uno) and an associated user interface (UI)
were developed in parallel [Figure 2(e)] that would enable the
digital configuration and control of the ventilation profile as
well as the ability to report anomalous events to the caregiver
through an ISO 60601-standardized alarm profile. As the
design progressed, extensive manufacturing and assembly
instructions were created [Figure 2(f)] that would enable others
to manufacture the VOV and would be continually updated
throughout the remainder of the project to reflect all design
modifications. A complete Institutional Animal Care and Use
Committee (IACUC) protocol was drafted and approved by
Vanderbilt in two days, enabling us to move forward with animal experiments.
Concept Refinement and Testing
The integration of the UI/embedded controller with version 3.0
led to the creation of version 3.1 [Figure 2(g)] on 2 April. Version 3.1 would be the first unit tested in an in vivo setting on
the next day. At our first live swine experiment on 3 April, we
observed insufficient gas exchange from our device, resulting in
the animal breathing out of synchronization with our ventilator.
This was found to be due to the existence of substantial dead
space in the ventilation circuit (specifics of which are provided

Protective Cover

User Interface
Ambu Bag
Receptacle

Pressure-Sensing
Single-Limb
Circuit

12-V Power
Supply

Tidal Volume
Adjustment Knob

Figure 1. The Vanderbilt Open Source Ventilator (VOV), version 4.0. The device is designed to compress a standard Ambu bag, which
is a widely available hand-squeezed device used to provide breathing support when transporting patients. The design features a
car windshield wiper motor, Arduino-based control, and the valves and sensors needed to effectively and safely provide mechanical
ventilation to COVID-19 patients.

MARCH 2021

IEEE ROBOTICS & AUTOMATION MAGAZINE

103

IEEE Robotics & Automation Magazine - March 2021

Table of Contents for the Digital Edition of IEEE Robotics & Automation Magazine - March 2021