RC Ornithopter — Flight Stabilization & Telemetry System

An electronics-focused RC ornithopter featuring a dual-crank flapping mechanism,
MPU6050-based flight stabilization, and real-time PID control.

Live PID Simulator · Circuit Design · Arduino Code · How It Works

Project Overview

This project is an electronics-domain implementation of a Radio-Controlled (RC) Ornithopter — a flapping-wing micro aerial vehicle (MAV). The mechanical structure uses a dual-crank gear mechanism to convert motor rotation into symmetrical wing flapping. On top of this, a custom Flight Stabilization and Telemetry System is designed and implemented.

Core Electronics Features

Feature	Implementation
Flight Controller	Arduino Nano / ESP32
Stabilization	MPU6050 IMU (6-axis Gyro + Accelerometer)
Control Algorithm	Hybrid RC + PID (Proportional-Integral-Derivative)
Sensor Fusion	Complementary Filter (gyro + accelerometer)
Actuation	BLDC Motor via ESC + 2× SG90 Servo
Remote Control	FlySky FS-i6 Transmitter / FS-iA6 Receiver
Power	2S/3S LiPo via ESC BEC

Mechanism: The Dual-Crank Flapping System

The ornithopter uses a compound gear train to translate the rotary motion of a BLDC motor into the reciprocating flapping motion of its wings.

How It Works

The BLDC motor drives a small 10-tooth pinion gear.
The pinion meshes with a 70-tooth spur gear, achieving a gear reduction of 7:1 — this reduces speed and multiplies torque.
Each large gear has an offset crank pin. A connecting rod links this pin to a wing lever.
As the gears rotate, the connecting rods push and pull the wing levers, producing the flapping stroke.
Two cranks placed 180° apart ensure symmetrical flapping, preventing roll instability.

Gear Train Summary

Gear	Teeth	Role
Pinion (A)	10	Input from motor
Spur Gear + Pinion (B+C)	70 + small	Intermediate stage
Driven Gear (D)	Large	Output shaft / crank

Why Dual-Crank?

Eliminates the dead-centre effect of single-crank designs
Ensures perfectly symmetrical flapping — the ornithopter flies straight
Reduces shock loads on the motor and ESC
Provides smoother power delivery and lower vibration

Electronics System Architecture

Components & Specifications

#	Component	Model	Key Specifications
1	BLDC Motor	F23700 (3500KV)	Composite material, high RPM, lightweight
2	ESC Controller	30A ESC	490Hz response, 16KHz motor freq, 2–3S LiPo, BEC 5V/3A
3	Servo Motor ×2	SG90	9g, 1.8 kgf·cm torque, 0.1s/60°, 4.8V
4	RC Transmitter	FlySky FS-i6	6CH, 2.4GHz AFHDS 2A, 4096 resolution
5	RC Receiver	FlySky FS-iA6	6CH, 2.4GHz, 500KHz bandwidth
6	IMU Sensor	MPU6050	6-axis, I2C, ±250°/s gyro, ±2g accel
7	Microcontroller	Arduino Nano	ATmega328P, 16MHz, 32KB flash
8	Battery	2S/3S LiPo	Lightweight, 45min operating time
9	FPV Camera	WiFi Mini Cam	640×480 VGA, M-JPEG, 10fps, 2.4G WiFi, 8g
10	Frame Material	Fiberglass Sheet	Lightweight, rigid structural support
11	Wing Membrane	Polyethylene Cover	Lightweight, flexible, aerofoil surface
12	Connectors	JST-XH Plugs	Standard RC/LiPo connectors

Circuit Design & Wiring

Power Architecture

Critical: NEVER power servos or the motor directly from the Arduino's 5V pin. Always use the ESC's built-in BEC (Battery Eliminator Circuit) for the 5V rail.

Pin Wiring Table

Component	Wire / Pin	Arduino Nano Pin
MPU6050	VCC	5V (from BEC)
MPU6050	GND	GND
MPU6050	SDA	A4
MPU6050	SCL	A5
RC Receiver CH3	Throttle Signal	D2 (PWM Input)
RC Receiver CH1	Yaw Signal	D3 (PWM Input)
RC Receiver CH2	Pitch Signal	D4 (PWM Input)
ESC Signal	PWM Input	D9 (PWM Output)
Servo 1 (Elevator)	Signal	D10 (PWM Output)
Servo 2 (Rudder)	Signal	D11 (PWM Output)

RC Stick Mapping (Mode 2 — Default)

Stick	Axis	Function	Maps To
Left Stick	Up/Down	Throttle	Flapping speed (ESC)
Left Stick	Left/Right	Yaw	Rudder servo
Right Stick	Up/Down	Pitch	Elevator servo (+ PID setpoint)
Right Stick	Left/Right	Roll	(Passive — wing geometry)

Arduino Flight Controller Code

Serial Monitor — Startup Output

The screenshot below shows the Arduino IDE Serial Monitor at 115200 baud during a calibration-to-flight sequence. The I²C bus scan confirms MPU6050 at address 0x68, followed by receiver channel verification and PID loop telemetry.

Observed: All three RC channels lock at ~1500 µs (center) at startup. PID output converges toward zero within 0.5 s as the complementary filter settles.

File: `src/flight_controller_basic.ino`

Manual RC passthrough — no stabilization. Used for Phase 1 testing.

#include <Servo.h>
#include <Wire.h>

const int escPin            = 9;
const int elevatorServoPin  = 10;
const int rudderServoPin    = 11;
const int throttleInPin     = 2;
const int pitchInPin        = 3;
const int yawInPin          = 4;

Servo flappingMotor, elevatorServo, rudderServo;

void setup() {
  Serial.begin(115200);
  Wire.begin();

  // Wake MPU6050
  Wire.beginTransmission(0x68);
  Wire.write(0x6B); Wire.write(0);
  Wire.endTransmission(true);

  flappingMotor.attach(escPin, 1000, 2000);
  elevatorServo.attach(elevatorServoPin);
  rudderServo.attach(rudderServoPin);

  // ESC arming
  flappingMotor.writeMicroseconds(1000);
  delay(2000);
  Serial.println("ESC Armed. Flight loop starting...");
}

void loop() {
  int thr = pulseIn(throttleInPin, HIGH, 25000);
  int pitch = pulseIn(pitchInPin, HIGH, 25000);
  int yaw   = pulseIn(yawInPin,   HIGH, 25000);

  thr   = constrain(thr, 1000, 2000);
  pitch = constrain(pitch, 1000, 2000);
  yaw   = constrain(yaw, 1000, 2000);

  flappingMotor.writeMicroseconds(thr);
  elevatorServo.writeMicroseconds(pitch);
  rudderServo.writeMicroseconds(yaw);

  delay(10); // ~100 Hz
}

File: `src/flight_controller_pid.ino`

Full PID + RC hybrid — active pitch stabilization.

#include <Servo.h>
#include <Wire.h>

// ── PID Gains (tuned empirically) ───────────────────────────
float Kp = 1.80, Ki = 0.05, Kd = 0.40;

// ── Complementary Filter ────────────────────────────────────
const float alpha = 0.98;
float pitch_angle = 0.0;

// ── PID state ───────────────────────────────────────────────
float pid_integral  = 0.0;
float prev_pitch_err = 0.0;
unsigned long prev_time;

// ── Raw IMU ─────────────────────────────────────────────────
int16_t ax, ay, az, gx, gy, gz;

float readPitch_accel() {
  return atan2(ay, az) * 180.0 / PI;
}

void loop() {
  // ── Read MPU6050 ──────────────────────────────────────────
  Wire.beginTransmission(0x68);
  Wire.write(0x3B); Wire.endTransmission(false);
  Wire.requestFrom(0x68, 14, true);
  ax = Wire.read()<<8 | Wire.read();
  ay = Wire.read()<<8 | Wire.read();
  az = Wire.read()<<8 | Wire.read();
  Wire.read(); Wire.read();  // temperature
  gx = Wire.read()<<8 | Wire.read();
  gy = Wire.read()<<8 | Wire.read();

  // ── Complementary Filter ─────────────────────────────────
  unsigned long now = micros();
  float dt = (now - prev_time) / 1e6;
  prev_time = now;

  float gyro_rate = gy / 131.0;           // °/s
  float accel_pitch = readPitch_accel();
  pitch_angle = alpha*(pitch_angle + gyro_rate*dt) + (1-alpha)*accel_pitch;

  // ── RC Pitch Setpoint ────────────────────────────────────
  int rc_pitch = pulseIn(pitchInPin, HIGH, 25000);
  float setpoint = map(rc_pitch, 1000, 2000, -15, 15);  // ° target

  // ── PID ──────────────────────────────────────────────────
  float error = setpoint - pitch_angle;
  pid_integral += error * dt;
  pid_integral  = constrain(pid_integral, -20, 20);
  float derivative = (error - prev_pitch_err) / dt;
  prev_pitch_err   = error;

  float pid_out = Kp*error + Ki*pid_integral + Kd*derivative;
  pid_out = constrain(pid_out, -30, 30);

  // ── Servo Output ─────────────────────────────────────────
  int elev_cmd = 1500 + (int)(pid_out * 6.0);
  elev_cmd = constrain(elev_cmd, 1000, 2000);
  elevatorServo.writeMicroseconds(elev_cmd);

  // ── Telemetry ────────────────────────────────────────────
  Serial.print("t="); Serial.print(millis()/1000.0, 2);
  Serial.print("s  Pitch:"); Serial.print(pitch_angle, 1);
  Serial.print("°  err:"); Serial.print(error, 2);
  Serial.print("  PID_out:"); Serial.println(pid_out, 2);

  delay(10);
}

PID Stabilization System

PID Step Response — Gain Comparison

The plot below was generated from flight log data replayed through the PID algorithm with different gain sets. The optimal PID (Kp=1.8, Ki=0.05, Kd=0.4) settles within ±0.8° in under 0.6 s with no steady-state error.

Gain Configuration	Behaviour	Settling Time
Kp=0.5, Ki=0, Kd=0	Sluggish, large offset	Never settles
Kp=4.0, Ki=0, Kd=0	Oscillates continuously	—
Kp=1.8, Ki=0, Kd=0.4	Fast, small residual error	~0.9 s
Kp=1.8, Ki=0.05, Kd=0.4	Optimal — zero steady-state	~0.6 s

PID Tuning Procedure

Step 1: Set all gains to zero.
Step 2: Raise Kp until the ornithopter oscillates around level → back off 20%.
Step 3: Raise Kd to dampen oscillations.
Step 4: Raise Ki slowly to eliminate any residual tilt offset.

Symptom	Corrective Action
Slow response to disturbance	Increase Kp
Persistent tilt offset	Increase Ki
Oscillating / shaking	Increase Kd or reduce Kp
Violent overcorrection	Reduce Kp, increase Kd

IMU Data & Sensor Fusion

MPU6050 In-Flight Capture

Raw gyroscope (Roll/Pitch rate in °/s) and accelerometer Z-axis data captured via Serial at 100 Hz during a 2-second flight window. The 5.2 Hz flapping oscillation is clearly visible in both axes.

The complementary filter (α = 0.98) blends:

Gyroscope — accurate short-term angular rate, but drifts over time
Accelerometer — absolute tilt reference, but noisy due to wing vibration

At α = 0.98, the gyroscope dominates at high frequencies (wing vibration rejection) while the accelerometer slowly corrects long-term drift.

Wing Beat Frequency Analysis

FFT — Accelerometer Z-axis (5-second capture)

Post-flight FFT analysis of the accelerometer Z-axis data confirms 5.2 Hz fundamental wing beat frequency, with visible 2nd (10.4 Hz) and 3rd (15.6 Hz) harmonics consistent with the dual-crank mechanism's kinematics.

Peak	Frequency	Amplitude	Interpretation
Fundamental	5.2 Hz	2.1 °	Primary wing stroke
2nd Harmonic	10.4 Hz	0.8 °	Gear mesh / crank return
3rd Harmonic	15.6 Hz	0.3 °	Structural mode

The PID derivative term (Kd) must be low-pass filtered to prevent these harmonics from coupling into the servo output.

Live PID Simulator

An interactive browser-based PID tuning simulator is hosted on GitHub Pages. Visualize the effect of Kp, Ki, Kd on simulated ornithopter pitch response before touching hardware.

** Open Live Simulator**

Features:

Real-time pitch response canvas
Adjustable Kp / Ki / Kd sliders
Wind gust disturbance injection button
Animated ornithopter responding to PID output
Complementary filter visualization toggle

Aerodynamic Calculations

Bernoulli Lift Equation

L = Cl × ρ × A × v² / 2

  L  = Lift force (N)
  Cl = Coefficient of Lift = 1.6
  ρ  = Air density = 1.2041 kg/m³
  A  = Wing area = 0.4032 m²
  v  = Forward velocity (m/s)

Aspect Ratio

AR = b² / S = (0.51)² / (0.4032) ≈ 0.645

→ Low aspect ratio — prioritizes maneuverability over glide efficiency,
  appropriate for a flapping-wing MAV.

Measured Flight Parameters

Parameter	Value
Total MAV Weight	98.2 g
Wing Span	510 mm
Wing Area	403.2 cm²
Flapping Amplitude	40°
Flapping Frequency	5.2 Hz
Lift Coefficient (Cl)	1.6
Gear Ratio	7:1
Cruising Throttle	~70%

Structural Safety Condition

σ_max ≤ σ_yt / n

  σ_yt = Yield strength of fiberglass frame
  n    = Factor of safety (≥ 2 recommended)

Test Results & Telemetry

Pitch PID Response — In-Flight Data

The graph below shows pitch angle (°) over a 3-second flight segment including a simulated wind gust disturbance at t = 0.3 s. The PID controller returns pitch to ±0.8° within 0.5 s.

Three-Phase Testing

Phase 1 — Manual RC Passthrough (No Stabilization)

Sustained flight achieved
Extremely difficult to control — pilot must constantly correct pitch
Frequent crashes due to flapping-induced pitch instability

Phase 2 — PD Controller on Elevator Servo

Pitch stabilization significantly improved
Gearbox rocker-link / shoulder identified as main failure point
Frame balance and gearbox mount revised

Phase 3 — Full PID + RC Hybrid

Semi-autonomous level flight demonstrated
~30 yards of controlled straight flight recorded
9 full wing-beat cycles captured per test window
ESC 30 A limit constrains maximum throttle to ~70%

Key Metrics

Metric	Value
Max sustained flight distance	~30 yards
Pitch error (steady state)	< ±0.8°
Pitch settling time (after disturbance)	~0.5 s
Wing beats per test window	9 cycles
PID loop rate	100 Hz
Serial telemetry rate	100 Hz @ 115200 baud

📁 Project Structure

How to Build

Prerequisites

Arduino IDE 2.x
Libraries: Servo.h (built-in), Wire.h (built-in)
FlySky FS-i6 transmitter bound to FS-iA6 receiver

Steps

Clone the repository

git clone https://github.com/KrHari09/RC-Ornithopter-Flight-Stabilization-Telemetry-System.git

Open the firmware
- Open src/flight_controller_pid.ino in Arduino IDE
Upload to Arduino Nano
- Board: Arduino Nano
- Processor: ATmega328P (Old Bootloader) ← required for clone boards
- Port: select your COM port
Calibrate ESC
- Power on with throttle at max → wait for beep sequence
- Lower throttle to minimum → wait for arming beep
- The delay(2000) in setup() handles this
Bind RC Receiver
- Follow FlySky FS-iA6 binding procedure
- Verify CH1/CH2/CH3 PWM signals on Serial Monitor (should read ~1500 µs at center)
Tune PID on bench
- Hold ornithopter, apply manual pitch disturbances
- Watch Serial Monitor for pitch angle and PID output
- Adjust Kp, Ki, Kd using the tuning table above
Flight test in open area
- Start at ~30% throttle
- Verify elevator responds to RC pitch input
- Verify PID corrects unintended pitch changes

References

Chronister, N. (1999). The Ornithopter Design Manual. The Ornithopter Zone.
Mueller, T. J. (2001). Fixed and Flapping Wing Aerodynamics for Micro Air Vehicle Applications. AIAA. ISBN 1-56347-517-0.
Azuma, A. (2006). The Biokinetics of Flying and Swimming (2nd ed.). AIAA. ISBN 1-56347-781-5.
DeLaurier, J. D. (1999). The Development and Testing of a Full-Scale Piloted Ornithopter. Canadian Aeronautics and Space Journal, 45(2), 72–82.
Warrick, D., Tobalske, B., Powers, D., & Dickinson, M. (2010). The Aerodynamics of Hummingbird Flight. AIAA.
Nagakiran, M., Kannakumar, J., & Salmon, A. (2019). Construction and Operation of Ornithopter Using RC. JETIR, Vol. 6, Issue 1. ISSN-2349-5162.

Author

Hari Kumar
B.Tech, Electronics and Communication Engineering, NIT Jamshedpur

Built with electronics, control theory, and a love for bio-inspired flight.

Name		Name	Last commit message	Last commit date
README.md		README.md
circuit_schematic		circuit_schematic
component_list		component_list
flight_controller_basic.ino		flight_controller_basic.ino
flight_controller_pid.ino		flight_controller_pid.ino
flight_log		flight_log
pid_tuning_guide		pid_tuning_guide
script.js		script.js
style		style

Latest commit

History

Folders and files

Repository files navigation