IEEE Robotics & Automation Magazine - March 2021 - 110

UI features (potentiometers, buttons, switches, and an LCD
screen) allow the physician to interact with the ventilator and
monitor the status of both the ventilator and the patient, as
illustrated in Figure 5(b).
Control Architecture
The controller is implemented in the form of a finite-state
machine (FSM), illustrated in Figure 5(c). An alarm manager object keeps track of various alarm conditions outside of the loop and reports them to the user through a
combination of a flashing LED, ringing buzzer, and
message displayed on the UI LCD. We also note that
new settings may be programmed while the ventilator is
actively ventilating, without breaking the main loop,
enabling the physician to modify ventilation parameters
during operation based on feedback from the pressuresensing, single-limb circuit. Patient safety was of paramount importance in every design decision of the
VOV, particularly in regards to barotrauma avoidance.
The VOV implements active barotrauma avoidance
by modulating the TV if the inspiratory pressure exceeds
predefined thresholds, as displayed experimentally in

BPM

40
30
20

Desired
Actual

I/E Ratio

0
0

1
0.8
0.6
0.4
0.2
0

80 100 120 140 160 180 200
Cycle Number
(a)

Desired
Actual

80 100 120 140 160 180 200
Cycle Number

Scaling Parameters

(b)
2
1.5
1
α
β

0.5
0
0

80 100 120 140 160 180 200
Cycle Number
(c)

Figure 7. The end-cycle proportional timing controller for various
BPM and I/E settings: block diagram of the (a) BPM versus
cycle number (commanded and actual), (b) I/E ratio versus
cycle number (commanded and actual), and (c) speed scaling
parameters a and b versus cycle number.

110

IEEE ROBOTICS & AUTOMATION MAGAZINE

MARCH 2021

Figure 6. This is detailed in the Supplementary File, as
are numerous other control-based and mechanical-based
safety solutions.
End-Cycle Proportional Timing Control
Methodology
Windshield wiper motors run in open loop, so to achieve
accurate respiratory-rate timing, we have implemented an
end-cycle proportional timing controller. We sense the
position of the motor at the two most important points in
the respiratory cycle (full inspiration and full expiration)
with a pair of limit switches, dead-reckon between these
two points, and adjust speeds on the next cycle as necessary to meet these respiratory timing requirements, based
on the error between the desired and actual inspiration/
expiration times. The specific hardware implementation of
the end-cycle proportional timing methodology is available in the Supplementary File. This proportional timing
update capability is demonstrated experimentally in Figure 7, where the BPM and I/E ratio were increased every
50 cycles (12 BPM at 1:4 I/E, 20 BPM at 1:3 I/E, 30 BPM
at 1:2 I/E, and 40 BPM at 1:1 I/E) while the VOV was
actively ventilating a test lung apparatus with a built-in
compliance of 20 mL/hPa. As can be observed, the VOV is
quick to converge to the new settings (within 30% of the
desired setting after a single breath cycle) and with negligible steady-state error.
In Vivo and In Vitro Testing
In preparation for the FDA EUA submission, the VOV
was experimentally validated using a combination of in
vitro validation in a calibrated mechanical test lung and
live animal testing using an anesthetized swine model.
In Vivo Swine Study
Two live animal studies were performed in which the
VOV provided continuous ventilation to an anesthetized
swine for four hours. In the first study, as mentioned in
the " VOV Development Process " section, there was
insufficient gas exchange due to the length of the ET tubing. For a more detailed discussion of this, see " Insights
from First In Vitro Swine Study " in the Supplementary
File. In the second swine study, we corrected the problem
with a pressure-sensing, single-limb circuit [Figure 8(a)].
The swine was ventilated continuously for four hours
(with average settings of 20 BPM and an I/E ratio of 1:2)
as per our approved IACUC protocol. Throughout the
course of the second experiment, the swine remained
hemodynamically normal, with adequate oxygenation,
ventilation, and a normal pH. Subsequent histology
results revealed well-preserved alveolar structural integrity with no evidence of barotrauma or atelectasis [Figure 8(b) and (c)] [25].
It is likely that humidification would be useful in
the future long-term (e.g., weeks) use of this ventilator
with human patients. We successfully accomplished

IEEE Robotics & Automation Magazine - March 2021

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