Smart nebulizer technology

Drug delivery that only happens when it can reach the lungs.

Conventional nebulizers fire continuously. Half the drug is lost to the air on every exhale. PulmoSense detects each breath and delivers aerosol only during inhalation — closed-loop control validated in-silico across four patient profiles.

See it live View results
98–99%
Inhale-phase delivery efficiency (PulmoSense)
~33%
Typical conventional nebulizer efficiency
4
Patient profiles validated
30ms
Nebulizer response time constant (piezo mesh)
Interactive simulation

Watch the difference in real time.

Switch patient profiles, adjust breathing parameters, and watch PulmoSense outperform a conventional nebulizer live. Efficiency numbers are anchored to validated in-silico results.

The problem

Conventional nebulizers waste most of their drug.

A standard jet or mesh nebulizer runs continuously throughout the breath cycle. During exhalation, the aerosol goes into the air instead of the airway. With a typical 1:2 inhale-to-exhale ratio, only one third of the session is useful.

Conventional (1:2 I:E)~67% wasted
Conventional (COPD 1:3 I:E)~75% wasted
PulmoSense (any profile)1–2% wasted

Waste percentages for conventional nebulizers are derived from the inhale fraction of the breath cycle. PulmoSense waste reflects the physical 30 ms piezo mesh decay tail validated in simulation.

~67%
of aerosol from a conventional nebulizer is exhaled into the room in a standard adult breathing pattern. For COPD patients, this rises to 75%.
3×
more drug is consumed in a conventional session to deliver the same dose to the lungs. PulmoSense cuts drug consumption proportionally — meaningful for expensive biologics and high-frequency treatments.
98–99%
inhale-phase delivery efficiency measured in PulmoSense in-silico validation across all four patient profiles, using validated per-profile PID gains and a Simulink closed-loop model.
How it works

Five components. One closed loop.

The control logic is validated in simulation using a Model-Based Design pipeline before any hardware is built. Every block maps to a real device subsystem.

Block A
Virtual Patient
Generates a realistic breathing waveform per patient profile. Half-sine inhale, exponential exhale. Maps to the SDP810 differential pressure sensor in hardware.
Breathing block
Block B
Phase Detector
Reads the flow signal and outputs a yes/no inhale flag. Includes predictive early-cutoff that shuts the neb before exhale starts — the key to 98–99% efficiency.
Phase block
Block C
PID Controller
Classic parallel PID with per-profile validated gains. Only active during the inhale window. Output is a duty cycle (0–100%) driving the mesh actuator.
PID block
Block D
Nebulizer Plant
Simulates the physical piezo mesh element with a 30 ms first-order lag. Captures the real ramp-up behavior instead of assuming instant response.
Neb + Gain
Block E
Gain Scheduler
Selects the correct validated PID gain set for the active patient profile. Per-profile gains are tuned in simulation and stored in a lookup table loaded at runtime — no re-tuning on the device.
NVM lookup
Patient Profile
→
Breathing waveform
→
Phase detection
→
PID (inhale only)
→
Saturation 0–100%
→
Neb plant 1/(0.03s+1)
→
Aerosol output
Validated results

In-silico performance across all four patient profiles.

Results from the Phase 1 in-silico validation report (v2.3). Per-profile PID gains tuned in simulation and validated in the closed-loop Simulink model. Bench testing on hardware is the next step.

Normal Adult
99.4%
Inhale-phase delivery
Kp 2.000
Ki 1.000
Kd 0.033
Pediatric
98.8%
Inhale-phase delivery
Kp 0.109
Ki 1.000
Kd 0.083
COPD
98.0%
Inhale-phase delivery
Kp 0.100
Ki 1.000
Kd 0.026
Asthma
98.0%
Inhale-phase delivery
Kp 0.100
Ki 1.000
Kd 0.016
These results are from Phase 1 in-silico pre-validation only. Per FDA Modernization Act 2.0 (2022) and ANVISA RDC 751/2022, computational simulation is accepted as pre-clinical validation evidence. Bench testing on physical hardware is required before any clinical or regulatory submission.
Team

Built at the intersection of engineering and clinical need.

Founder & Engineer
Fahad Alhuthaifi
PulmoSense LLC
M.Eng. Biomedical Engineering, Cornell University (2025). Background in clinical and process engineering. Designed the MBD pipeline, closed-loop PID control architecture, and Simulink in-silico validation model.
Main Advisor & Academic Partner
Prof. Suélia S. R. Fleury Rosa
University of Brasília · Cornell University
Associate Professor of Biomedical Engineering at UnB with a postdoctorate from MIT Media Lab. Expert in physiological systems modeling, in-silico simulation of biological systems, and medical device development through Phase I–IV clinical trials. Main academic advisor on PulmoSense, providing expertise in translational health research and regulatory pathway to ANVISA and CONEP in Brazil.
Clinical & Brazil Partner
BioEny
Brazil · Ana Karoline, Director of Tissue Engineering
Clinical development and Brazil market partner. Collaborating on manuscript, patient data, and pathway to first-in-human validation in the Brazilian regulatory framework.
Contact

Interested in partnering or investing?

PulmoSense is at the in-silico pre-validation stage. We are currently seeking partners for PCB design and fabrication (BOM and Gerber phase) and conversations with strategic investors in respiratory drug delivery.

Get in touch Explore the simulation