Drone Anenometer
A primary difficulty that pilots face is weather conditions that endanger aircraft control or integrity.
Although some weather hazards can be anticipated and minimized by checking local forecasts and METAR prior to takeoff, windshear, downbursts, and temperature inversions and other localized invisible weather hazards can form and dissipate rapidly with no prior warning.
In manned aircraft, the pilot is immediately made aware of the effects of surrounding hazards.
However, with unmanned vehicles, the pilot’s distance from their aircraft makes it difficult for them to quickly perceive and react to invisible weather hazards immediately surrounding their aircraft. This can result in unpredictable aircraft behavior, crashes, and loss of RC control.
How can we inform UAV pilots of invisible localized weather conditions?
Existing Sensors
Commercially available anenometers fall into two categories:
A combo of rotating cups to monitor wind speed and a wind vane to determine wind direction
Ultrasonic sensor and transducer sensor array
Cupola design
Ultrasonic-transducer array
Drawbacks
Although the cup-and-vane design is generally affordable [1] and lightweight, it is too bulky to fit onboard a small aircraft. Additionally, it does not respond quickly to sudden wind changes due to inertia of the rotating parts.
The ultrasonic array, in contrast, has superior accuracy as it is unaffected by inertia. However, it is heavier and considerably more expensive [2] than the cup-and-vane design.
A New Anenometer
While interning at Ocean Alliance Inc., my manager Iain Kerr asked me to find a way to detect sudden crosswinds, which were blowing Snotbot off course while landing or collecting whale blow.
I designed a wireless thermal anemometer sensor package.
The anenometer uses 4 Rev. P Wind Sensors from ModernDevice, a MMA8451 accelerometer from Adafruit, two Arduino Nano, and 2 HC-12 wireless antennas, costing approximately $130 USD total.
By combining the relative readings from each anenometer, and subtracting the expected wind from the quadcopter's current motion, the Arduino is able to determine the relative direction and strength of the wind.
It is about 4 cubic inches in volume, not counting any mast or antennae to distance the sensor from the propellors, which is considerably smaller than either the cupola and wind vane design or the ultrasound design. The mast is included to distance the thermal sensors from the turbulent air surrounding the propellers.
Drawbacks include...
A sensitivity to air pressure (and thus, humidity and slow drift due to pressure front changes)
High power draw
Difficulty of calibration (propeller wash affects local wind readings, nonlinear thermal-windspeed relationship)
The anenometer mounted on Snotbot
The ground station display for receiving wireless wind strength readouts
Public Project Artifacts
Windspeed Equation
Circuit Diagram
References
[1] Rotating anenometers cost between $40-600 per unit, depending on size and materials.
[2] Sonic anenometers cost between $500-4000 per unit, depending on sensor quality.