Anish Laha/Autonomous VTOL TailsitterSHEET 02 / 12 ›
DWGAL-02

Autonomous VTOL Tailsitter

A two-motor flying wing that takes off on its tail, pitches over into wing-borne flight and flies autonomous survey missions. I chose the configuration, then modelled, analysed, printed, wired, tuned and flew it.

Role
Designer and builder, solo
When
Aug 2025 to May 2026
Context
IB Design Technology project, Bangkok
Tools
Fusion 360FEA and CFDCuraFDM printingPixhawkPX4ArduPilotMission PlannerMAVLink
Status
Flown: autonomous survey missions
2.5 kgMass, against a 3 kg limit
16 / 25Unique parts / printed parts
~120 hPrint time
8,167 THBBill of materials, against a 10,000 THB budget
4 to 5 acresMapped per flight
Flight test. Armed standing on its tail legs, it climbs on thrust alone, then pitches over and speeds up as the wing takes the weight.

The brief

Small Thai rice farms can't see water stress or pests until it shows from the ground, so farmers over-water and over-fertilise to be safe. Three requirements shaped the aircraft.

RequirementWhyWhat it forced
Map 3 to 5 ha in one flightThat's a typical farmA wing, for range
No runwayPaddy fields don't have oneVertical take-off and landing
No trained pilotNobody on the farm fliesFully autonomous missions
Under 3 kg and under 10,000 THBWeight and cost limits in the briefPrinted airframe, two motors, hobby-grade electronics

Choosing the configuration

I didn't start with a drone. I sketched ten concepts, from a rover to a floating boat, and scored each against a 23-point specification. The top three got physical prototypes, tested on grass, mud and slopes, checked in FEA and re-scored.

Quadruped, 14 / 23. Flat feet slipped in mud.
Quadruped, 14 / 23. Flat feet slipped in mud.
Rover, 15 / 23. Frame dragged in long grass.
Rover, 15 / 23. Frame dragged in long grass.
VTOL, 19 / 23. Never touches the field.
VTOL, 19 / 23. Never touches the field.

Why a tailsitter

ConfigurationNo runwayCruises on a wingMotors
MultirotorYesNo4
Fixed-wingNoYes1
QuadplaneYesYes4 + 1
Tailsitter (mine)YesYes2

A quadplane does both jobs but carries four lift motors as dead weight in cruise. A tailsitter uses the same two motors and two elevons for everything. The worked example below is why that matters for range.

Hover: T = mg = 2.5 × 9.81 ≈ 24.5 N Cruise: T = D = mg / (L/D) ≈ 24.5 / 8 ≈ 3 N

L/D ≈ 8 is an assumed value for a small flying wing. That's roughly 8 times less thrust once the wing carries the weight. The catch is that the hard part moves into the transition.

Standing on its tail legs, nose up
Standing on its tail legs, nose up
In wing-borne flight (render)
In wing-borne flight (render)

Airframe in Fusion 360

Exploded view: segmented fuselage, slide-on wing panels, motor mounts, elevons and tail legs
Exploded view: segmented fuselage, slide-on wing panels, motor mounts, elevons and tail legs
Orthographic drawing of the wing root panel
Orthographic drawing of the wing root panel

Analysis, and what it changed

CFD: pressure over the wing at 15 m/s
CFD: pressure over the wing at 15 m/s
FEA: hard-landing load, stress at the root joint
FEA: hard-landing load, stress at the root joint
Test and setupFindingChange
Wing root FEA: lift and thrust at the tips, motor base fixedStress peaks at the root, past what PLA toleratesRibs, fillets, root walls 30% thicker
Belly FEA: 20 N hard-landing loadPeak at the base plateTPU feet, dual-wall belly
CFD: airflow at 15 m/s cruiseClean pressure split, tip vorticesUpturned tips, TPU leading edge
Bench: 0.5 m drops, 30 elevon hinge cyclesDents on hard soil; hinge whitening after 25 cyclesRubber pads, thicker hinge, TPU insert

Looking back, the FEA peaks sit at the fixed constraints, so they're partly an artefact of how I held the model. A mesh and boundary-condition check is the next step I'd take before trusting those numbers.

Printing and assembly

Early wing panel print with spar socket
Early wing panel print with spar socket
Wing panels on the carbon spars
Wing panels on the carbon spars
Dry fit before final assembly
Dry fit before final assembly

Electronics and flight stack

My wiring diagram for the build: Pixhawk, ESCs and motors, elevon servos, RC receiver, GPS and power
My wiring diagram for the build: Pixhawk, ESCs and motors, elevon servos, RC receiver, GPS and power
LayerWhat I used
Flight controllerPixhawk running PX4 on NuttX. It has also flown on ArduPilot.
SensingIMU, barometer, GPS and compass
EstimationEKF2 state estimator
ControlMulticopter, fixed-wing and transition controllers
Actuation2 motors on ESCs, 2 elevon servos
SafetyRC override and kill switch
Ground stationMission Planner over MAVLink
PowerLiPo through a power module

Layout and starting parameters referenced the open-source eXplora tailsitter. The airframe is my own design.

Bring-up, in the order I did it

Pixhawk and harness going into the bay
Pixhawk and harness going into the bay
  1. Mounting. Flight controller on foam pads so motor vibration doesn't corrupt the IMU. GPS on a mast away from the power wires. Camera on silicone dampers.
  2. Power. Soldered harness, heat-shrink, polarity checked on a multimeter before plugging anything in.
  3. Firmware. PX4 flashed, tailsitter airframe selected.
  4. Calibration. Accelerometer, gyro, compass and radio.
  5. Actuators. Props off: motor order, spin direction, elevon throws. A reversed elevon flips a tailsitter on take-off.
  6. Safety. Kill switch and failsafes tested at low throttle.
  7. Hover tuning. Rate PIDs first, then attitude.
  8. Transition. Transitions at height, then full missions.

Control

Control is a cascade. The mission gives a waypoint, position error becomes a velocity setpoint, velocity becomes an attitude, and attitude becomes body rates. EKF2's estimate feeds back into every loop. I tuned from the inside out: rate loops first, then attitude.

LoopControllerOutput
MissionNext waypointPosition setpoint
PositionPVelocity setpoint
VelocityPIDAttitude setpoint
AttitudePRate setpoint
Body ratesPIDMoments
AllocationMixerMotors and elevons

The actuator mapping is the same in hover and cruise. What changes is how each motion looks from the ground.

Body axisActuatorIn hover it looks likeIn cruise
Along the fuselageElevons oppositeSpinning on the spotRoll
Along the spanElevons togetherTilting fore and aftPitch
Out of the topDifferential thrustTilting sidewaysYaw

The elevons only have authority in hover because they sit in the propwash.

Transition as a mode switch

Schematic animation of the whole mission: vertical take-off, forward transition, fixed-wing survey lanes, back-transition and vertical landing. The readouts are illustrative.
PX4 parameterDefaultRole
VT_ARSP_BLEND8 m/sStart blending multicopter and fixed-wing control
VT_ARSP_TRANS10 m/sSwitch to fixed-wing
VT_TRANS_TIMEOUT15 sAbort if the transition is slow
VT_B_DEC_MSS2 m/s²Back-transition deceleration
VT_FW_QC_P / _RoffQuad-chute: drop back to hover if pitch or roll pass a limit

The back-transition matters most for planning, because the aircraft needs room to slow down.

run-out = v² / 2a = 15² / (2 × 2) ≈ 56 m

So the mission has to start the back-transition about 56 m before the pad, on a straight, clear line.

Autonomous survey missions

SeqMAVLink commandWhat it commits to
0NAV_VTOL_TAKEOFFClimb, then transition
1DO_SET_CAM_TRIGG_DISTA photo every N metres
2 to 13NAV_WAYPOINTLawnmower survey lanes
14NAV_VTOL_LANDBack-transition and land
TRY ITPlan a survey. Click to drop waypoints, then fly it. The orange frames are photos.

Sizing the survey grid

footprint w = 2h · tan(FOV / 2) lane gap d = w · (1 − sidelap) h = 60 m, FOV = 60°, sidelap 70% → w ≈ 69 m, d ≈ 21 m

A 150 × 120 m field (about 4.5 acres) is 6 lanes and about 1 km of track, roughly 70 s at 15 m/s. That matches the 4 to 5 acres I covered per flight.

Crop-health sensing

The payload estimates crop health with NDVI. Chlorophyll absorbs red light and healthy leaf structure reflects near-infrared, so a stressed plant reflects less NIR and more red, and its NDVI falls.

NDVI = (NIR − Red) / (NIR + Red) Healthy: (0.50 − 0.08) / (0.50 + 0.08) = 0.72 Stressed: (0.30 − 0.12) / (0.30 + 0.12) = 0.43
Printed stress cards for the camera test
Printed stress cards for the camera test
Dense vegetation card
Dense vegetation card

My version used a camera and threshold bands on the microcontroller. The honest limit is that an RGB camera only approximates NDVI. A real measurement needs a near-infrared band.

The farmer app I designed: icon-led, Thai text
The farmer app I designed: icon-led, Thai text
One-tap emergency stop
One-tap emergency stop

Test results, and what failed

TestResultWhat changed
Crop-stress cards (NDVI)4 of 5 right. The miss was at NDVI 0.41, right on my threshold.Recalibrate thresholds
Kill switchMotors cut immediately, 3 of 3 trialsNothing needed
Weight2.5 kg against a 3 kg limitNothing needed
Five untrained users4 of 5 armed it and set a mode in under 2 minClearer start-up cues
App icons43 of 50 read correctlyNothing needed
Powered VTOL by the school deadlineNot flown in time. I rated it 1 of 5.Flown since: autonomous survey missions, terrain mapped
X 0000 Y 0000 HDG 000 ALT 35000 FT