Final Project | Sleep Monitoring

A compact wearable that detects snoring, tracks sleep position, and monitors temperature and humidity. The data are processed on the device and shown on the OLED and a live web dashboard.


Author	Mohammadreza Omidali
FabLab	Oulu Super FabLab, University of Oulu
Microcontroller	Seeed XIAO ESP32-S3
Sensors	ICM-20948 IMU, DHT22 temperature and humidity sensor, and ICS-43434 I2S microphone
Output	0.96-inch I2C OLED and WiFi web dashboard
Power	3.7 V, 600 mAh LiPo battery
Enclosure and mounting	Two-part snap-fit PLA enclosure, UV-printed lid, and elastic chest band
License	CC BY-NC-ND 4.0
Repository	Fab Academy 2026, Oulu Super FabLab, Mohammadreza Omidali

Presentation

Slide

Download slide: presentation.png

Video

Watch the final presentation video:

▶ Presentation Video (MP4)

1. The Idea

My first idea was to work with a high frequency time-of-flight ultrasound sensor for biomedical applications. Because a suitable sensor was not available, I changed the project plan.

Then I decided to make a small sleep-monitoring device with low-cost that could monitor snoring, sleeping position, temperature, and humidity during sleep.

Prior work and inspiration

I reviewed some related projects in Week 17.

2. Answering the Official Questions

I answered the official final project questions in more detail in Week 17.

3. System Overview

The main part of the device is a XIAO ESP32-S3. It reads the three sensors, updates the OLED, and sends the data to a web dashboard through WiFi. The system does not need an app, cloud service, or external server.

I showed and explained the block diagram in Week 15.

Layer	Protocol	Components
Motion sensing	I2C	ICM-20948
Environmental sensing	Digital single-wire	DHT22
Audio sensing	I2S	ICS-43434 microphone
Local output	I2C	SSD1306 OLED
Connectivity	WiFi and WebSocket	Browser dashboard

4. Designs: 2D & 3D

The enclosure was designed and improved over several weeks:

The first sketches and 3D model were created in Week 02.
The component layout and system integration were developed in Week 15.
The lid graphics were prepared in Inkscape and UV printed in Week 16.

Enclosure

I designed the enclosure in Fusion 360 after measuring the real components. I first made a larger box and then reduced its size after placing the PCB, battery, sensors, and display in the model.

The enclosure has:

a two-part snap-fit body and lid
internal supports for the PCB
a space under the PCB for the LiPo battery
openings for the OLED, microphone, and sensors
an attachment point for the chest band
a UV-printed lid with the project name and logos

The enclosure was printed from PLA on a Prusa CORE One L.

5. Fabrication

5.1 3D Printing

The enclosure body and lid were sliced in PrusaSlicer and printed on a Prusa CORE One L.

Setting	Value
Material	PLA
Layer height	0.15 mm
Infill	20%
Supports	No

5.2 Laser Cutting

Before making the final PCB, I made a temporary holder from plywood with the laser cutter. I used it to place the components and check their arrangement inside the enclosure.

I used the tested plywood preset available at Oulu Super FabLab. I checked the focus and position before cutting. I did not save the exact power and speed values because this holder was used only during development. I do not want to add values that I did not record.

Item	Details
Material	Plywood
Purpose	Temporary component holder
Design software	Fusion 360 and Inkscape
Cutting settings	Material preset available at Oulu Super FabLab

The exact numerical power and speed settings were not saved. The general laser characterization, including power, speed, focus, and kerf tests, is documented in Week 03.

5.3 UV Printing

The lid was 3D printed from PLA. It was not cut with a UV laser. I used the Mimaki UJF-3042MkII e flatbed UV inkjet printer to print the project name and logos on its surface. Because this machine prints UV-curable ink, laser power, cutting speed, and number of passes do not apply.

First, I printed an outline to find the correct position of the lid. Then I placed the lid inside the outline and printed the logos and text.

The first print did not have good quality because the logo files had unwanted backgrounds. I downloaded the original logo files, edited them in Inkscape, and printed them again. The second result was much clearer.

Item	Details
Machine	Mimaki UJF-3042MkII e
Object	3D-printed PLA enclosure lid
Lid height	4 mm
Media thickness setting	5 mm
Software	Mimaki RasterLink7
Printing method	Direct UV inkjet printing

The photos below show the first and final UV-printing results. The complete process is documented in Week 16.

6. Electronics

6.1 Design

I designed the schematic and PCB in KiCad. I added and connected the components one by one.

First, I selected the symbols for the XIAO ESP32-S3 and the sensors. I then connected the components in the schematic. The images below show the component selection and the completed schematic.

After finishing the schematic, I moved the components to the PCB Editor and routed the tracks. I designed a one-layer PCB and added a ground plane. I checked the schematic with ERC and the PCB with DRC before fabrication.

The OLED footprint was not available in the default KiCad libraries, so I added a footprint from this repository.

The next images show the final PCB layout and its 3D view. The 3D view helped me check the component positions before milling the board.

6.2 Production

I exported the copper and board-outline files from KiCad and checked them in Gerber Viewer. I then placed an FR4 copper board in the LPKF milling machine. The first two images show the board during milling and the copper tracks after milling.

After milling, I removed the board and checked the tracks and board outline. I also used the microscope to see the small details more clearly.

Next, I cleaned the board and started soldering the components. I used the microscope during soldering because the XIAO pins and sensor connections were small.

I prepared the required openings and checked the connections around the board. I inspected the soldered areas closely before continuing.

I soldered the XIAO ESP32-S3 and connected the battery wires. I checked that the pins were connected correctly and that there were no solder bridges.

After the main soldering was finished, I added the remaining connectors and prepared the board for installation inside the enclosure.

Finally, I mounted the PCB inside the enclosure and connected the sensors, OLED, and battery. The last images show the board fixed inside the enclosure and the main components placed together.

6.3 Wiring Reference

I used following datasheets during wiring:

Component	Signal	GPIO	XIAO pin	Notes
DHT22	DATA	GPIO4	D3	Digital input
ICM-20948	SDA	GPIO5	SDA / D4	Shared I2C bus
ICM-20948	SCL	GPIO6	SCL / D5	Shared I2C bus
OLED	SDA	GPIO5	SDA / D4	Address 0x3C
OLED	SCL	GPIO6	SCL / D5	Shared I2C bus
ICS-43434	BCLK	GPIO43	TX / D6	I2S clock
ICS-43434	WS	GPIO44	RX / D7	I2S word select
ICS-43434	DOUT	GPIO2	D1	I2S data
ICS-43434	L/R	GND	GND	Left channel
All components	VCC	3.3 V	3.3 V	Do not connect to 5 V
All components	GND	GND	GND	Common ground

The temporary holder was tested before the final PCB was produced.

7. Embedded Programming

The XIAO ESP32-S3 reads all three sensors and updates the OLED and web dashboard.

Sleep position

The ICM-20948 provides accelerometer and gyroscope data. Roll and pitch are calculated with a complementary filter that combines 96% gyroscope data and 4% accelerometer data. This reduces the drift that happens when only the gyroscope is used. The final position is shown as back, left side, or right side.

Snore detection

The ICS-43434 microphone sends digital audio samples to the ESP32-S3 through I2S. The program processes the sound level on the device and uses a threshold to mark a possible snoring event. Raw audio is not sent to the dashboard.

Temperature and humidity

The DHT22 provides room temperature and relative humidity. These values are shown on the OLED and sent to the dashboard.

8. Networking & Interface

The ESP32-S3 provides the web page directly. After connecting to the device through WiFi, the user opens the local address in a browser. No application installation is needed.

The dashboard uses WebSocket communication, so new readings can be displayed without refreshing the page. The interface shows temperature, humidity, sleep position, and snore status. Start and Stop controls allow the user to control the data stream. In the interface test, sensor values were sent every two seconds.

The dashboard code combines:

HTML for the page structure
CSS for the layout and colours
JavaScript for live updates and controls
C++ code on the ESP32-S3 for sensor reading and WebSocket communication

The interface development is documented in Week 14.

9. System Integration & Assembly

The battery is placed below the PCB holder, while the PCB, sensors, and OLED are placed in the main enclosure. The OLED and sensors are aligned with their openings in the lid. Vertical sockets were used to keep the inside organized.

The complete integration process is documented in Week 15.

10. Testing and Results

I tested each main part during development and again after integration. The current project is a functional engineering prototype. It is not a medical device and has not been clinically validated.

Test	Method	Result	Status
PCB operation	Uploaded simple test programs after milling and soldering	The microcontroller and connected components operated correctly	Completed
DHT22 response	Moved the sensor near a warm laptop fan and observed the output	The reported temperature increased and the dashboard updated	Functional test completed; accuracy not calibrated
IMU response	Moved the sensor through different orientations and observed roll and pitch	The values changed with sensor orientation and supported basic position classification	Functional test completed; classification accuracy not quantified
Snore detection	Tested microphone input with sound and silence	The microphone input and threshold-based event output responded	Functional test completed; detection accuracy not quantified
OLED output	Displayed device and sensor information	The OLED showed the programmed output	Completed
WiFi and WebSocket	Opened the local dashboard and streamed readings to a browser	The browser received live updates without page refresh	Short functional test completed; overnight stability not measured
Battery operation	Powered the assembled device with the 600 mAh LiPo battery	Portable battery operation was confirmed	Full-night runtime not measured
Enclosure fit	Installed the PCB, sensors, display, and battery and closed the lid	The components fitted inside and the snap-fit lid closed	Completed
Chest-band attachment	Attached the enclosure to the band	The device could be worn on the chest	Completed; comfort not studied overnight
UV-printed lid	Inspected the second print after replacing low-quality logos	The vector logos and text printed clearly	Completed

Current limitations

The device has not been compared with a clinical sleep monitor.
Snore-detection accuracy has not been measured over several nights.
Sleep-position classification has not been compared with annotated video.
Full-night battery life and WiFi stability have not been measured.
Comfort and usability have not been studied with participants.

These tests are planned for the next development stage and are not claimed as completed results in this project.

11. Bill of Materials

The table below lists the main materials and components used in the final project.

Item	Quantity	Use in the project
XIAO ESP32-S3	1	Main microcontroller and WiFi connection
ICM-20948	1	Motion and sleep-position sensing
DHT22	1	Temperature and humidity measurement
ICS-43434 I2S microphone	1	Sound and snoring input
SSD1306 0.96-inch OLED	1	Local display
3.7 V, 600 mAh LiPo battery	1	Portable power supply
FR4 copper board	1	Milled custom PCB
Pin headers, connectors, and wires	As needed	Electrical connections
PLA filament	About 50 g	3D-printed enclosure and lid
Plywood	1 small piece	Temporary laser-cut component holder, not in the final device but during the process
UV-curable ink	As needed	Logos and text printed on the lid

Most of the materials and components were available at Oulu Super FabLab. The component costs are listed in Week 17.

12. Fab Academy Skills Map

Fab Academy week	Application in this project
Week 02: Computer-aided design	Enclosure sketches and 3D design
Week 03: Computer-controlled cutting	Laser-cut temporary holder
Week 04: Embedded programming	Microcontroller programming
Week 05: 3D scanning and printing	PLA enclosure and lid
Week 06: Electronics design	KiCad schematic and PCB layout
Week 08: Electronics production	PCB milling, soldering, and testing
Week 09: Input devices	IMU, microphone, and DHT22
Week 10: Output devices	OLED display
Week 11: Networking and communications	WiFi and data communication
Week 13: Molding and casting	Not used in the final project BUT I do use it in my current work a lot
Week 14: Interface programming	Browser dashboard
Week 15: System integration	Packaging and assembly
Week 16: Wildcard	UV printing on the lid

13. Dissemination and Future Opportunities

The files and documentation for this Fab Academy prototype are shared under the license stated below. The next development stage is separate from this project.

For the next step, I would like to add a physiological sensor, such as an SpO2 or heart-rate sensor, and compare the device with a reference sleep monitor.

Before publishing new PCB designs, sensor combinations, or algorithms, I plan to discuss them with the University of Oulu Innovation Centre and an IP expert.

The dissemination and development plan is described in Week 18.

14. License

The current Fab Academy documentation is published under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 license.

This license allows people to copy and share the published material for non-commercial purposes with attribution, but it does not allow distribution of modified versions. Because of these restrictions, I describe the current project as source-available, not fully open-source.

The license and IP plan for future PoC development will be reviewed separately with the University of Oulu Innovation Centre before new technical details are published.

15. Reflection

This final project brought together almost everything I learned during Fab Academy. I worked with 2D and 3D design, laser cutting, 3D printing, UV printing, PCB design and production, sensors, embedded programming, networking, system integration, and a web interface. At the beginning, these were separate skills, but during this project I learned how they work together in one system.

One of the most difficult parts was fitting all the components inside a small enclosure. I first designed a temporary holder and cut it from plywood with the laser cutter. I used it to check the position of the real components. I also had problems during PCB production, soldering, sensor wiring, and UV printing. Small mistakes, such as a wrong pin number or a low-quality logo file, caused delays. These problems taught me to test each part separately before putting everything together.

Overall, I am happy with the final result. I built the project from the first sketch to an assembled prototype with a custom PCB, enclosure, sensors, and browser dashboard. I also learned that making a working prototype is only the first step and that more testing is needed before it can become a reliable product.

16. Acknowledgements

I would like to express my sincere gratitude to:

Jani Ylioja, the director of Oulu super FabLab, for his support and guidance during my FabAcademy journey.
Antti Mäntyniemi, the manager of Oulu super FabLab, everyday available and helping.
The all instructors at Oulu Super FabLab for their valuable assistance and expertise. Thank you so much to all of them.
Neil Gershenfeld for creating this exceptional course and fostering a global network community.
The FabAcademy program for providing the opportunity, resources, and collaborative environment that made this work possible.

17. Files I Created

I collected the main design and code files here so they can be downloaded directly.