Team Members

The advancement of 21st century technology had provided humans with machine learning (ML), long range telemetry, and machine automation (MA).  However, those innovations were not fully integrated into the human daily activities.  In the case of turbidity measurement, it is currently measured by manually collecting data using a probe or continuously collecting data using an immobile sensor system.  Both methods failed to provide an accurate turbidity level of a large body of water at different depths and locations.

Loch-Ness, an automated turbidity measuring drone, eliminates the aforementioned flaws using ML, long range telemetry, and MA. The drone is an unmanned underwater vehicle (UUV) equipped with a backscatter turbidity sensor, an acoustic total suspended solid (TSS) sensor, and a temperature sensor. The drone automatically moves in x, y, z planes using a GPS guided system via a long range telemetry transmission.  It collects data at different depths and locations then sends them to a cloud server.  The data is automatically converted into an interactive 3D interface of the body of water using a ML program. The interface can be access from any platforms (mobile, desktop, and laptop).  The drone has extremely low power consumption, hence, a combination of wind and solar generators is sufficient.      

Design & Development


Solar Panel - Monocrystalline solar panel by Renogy Wanderer. Output voltage: 12V. Output power: 400W. Solar efficiency: 21%. 42.2” x 19.6” x 1.38”.
Wind/Water flow Turbine - Waterlily Turbine. Output voltage: 12V. Output power: depends on wind/water speed. Charging initiation speed: 0.28m/s (0.62 mph). 8.2” x 8.2” x 4.7”.
Rechargeable Battery - (2x) MaxAmps Lipo in parallel. Capacity: 11,000 mAh/per. Voltage: 11.1V/per. True rating: 40C. Charge rating: 5C. Waterproof.
Charging Module - Charge Controller by Renogy Wanderer Li. Max solar current: 30A. Input voltage: 12 – 25V. Consumption current: 10mA.
Telemetry - Pixhawk 4 Telemetry. Input voltage: 5V DC. Transmit current: 100mA
Receive current: 25mA. Radio frequency/range: 915Mhz/+15 km.

Power Management Module - Pixhawk 4 PM Board. Output voltage: 5.1V @ 3A. Max current to ESC: 120A. Input voltage: 7 – 51V.
Movement Controller - Pixhawk 4. Includes 32-bit cortex processor, GPS/compass, magnetometer, barometer, gyroscope (2), and accelerometer (2). Input voltage: 4.75 – 5.25V.
ESC - (4x) Flycolor Fairly ESC. ESC cont. current/peak current: 20A/30A. ESC input voltage: 7.4 – 14.8V. ESC BEC output: 5V/1A
Servo Motors - (4x) Flycolor Fairly Motor. Motor speed: 2300 RPM/V. Max motor cont. current: 16A. Max motor cont. power: 270W.
Acoustic TSS /Temperature Sensor - Peacock UVP by Ubertone. Log data via Ubertone cloud software. Power consumption: 0.5 – 1W. Input voltage: 5V DC. Current drain: 100 – 200mA.

Turbidity Sensor (optical 850 nm) - OBS-3+ by Campbell Scientific. Log data via Campbell cloud software. Output voltage: 0-5V. Input voltage: 5-15V DC. Current drain: 17mA.

UUV frame - Polyethylene pipe. Floatable. Inert.

• Schematic
• Model
• Functionality
• The whole UUV is coated with hydrophobic spray (Neverwet) to prevent bio-growth and water entering.
• The PixHawk power management enters sleep-mode after completing a measurement routine to conserve power.
• During sleep-mode, the batteries are continuously charging from solar panels and turbine.
• The drone frame is composed of polyethylene pipes which keep the drone automatically floating on the water to collect sunlight or wind flow.
• Two ESC/motors are used to move in the x and y axis.
• Apply more current to one motor over the other to turn
• Apply same current to both motors to move forward
• The other two ESC/motors are used for diving and hovering at certain depth to collect data.
• Apply less or more current to both motors to either hover or dive (depending on the buoyance force).
• The information from the two sensors and the telemetry is collected and processed at a remoted office using Python platform (Keras + Tensorflow) to recreate an interactive 3-D graphical figure of the body of water.
• The graph includes location (x, y, and z), temperature, turbidity level, and TSS level.
• Water maintainers can click on any points see the data at the comfort of their office chair or any mobile devices.
• The TSS level and temperature data are used to compensate for the noise from the turbidity measurement.
• This confirmation technique results in the most accurate turbidity measurement.
• The MA program is transmitted to the UUV via the long range telemetry (+15 km) at 915 Mhz (pass through materials easier comparing to LoRa 2.4 Ghz).
• The 1 hectare (10,000 m2) is divided into 100 squares (10m x 10m).
• There are four UUV; each UUV is going to measure 50 squares at 5 different depths (0m, 3m, 6m, 9m, 12m).
• The UUV will be positioned and guided in the following pathway.
• The pathway allows each square to be measured twice (day/night).
• Each UUV has 1,728 sec (28.8 mins) at one square.
• The UUV will take 10 minutes or less to complete 5 measurements and automatically float to surface (turn off diving ESC/motors).
• The rest of the time, it will enter sleep-mode/charging while moving to the next square.


I was inspired by my frustration toward the government (in Vietnam) for incapable of measuring the turbidity level from construction site run-off. They measured the turbidity with a small probe. The probe measured a low turbidity level at the surface of the upstream while the bottom of the downstream was extremely cloudy. My family owns a fish farm in Vietnam, and the error in the government turbidity regulation has caused the fishes to stress and die. We have lost a lot of money because of turbidity. I joined this challenge to end this malpractice. I hope Keysight will give me a chance to solve this problem. I have no experiment with making video; therefore, it is not as good as I wanted it. However, I have detailed my plan in the description section. All of the components are listed. It is possible to bring Loch-Ness to live. Please help me do it.


Voting is closed!