Project Icarus

A quadcopter flight dynamics simulator built from scratch in Python, focused on controls, state estimation, and autonomy. The verified GNC logic is being ported to an embedded C++ architecture targeting a CUAV V5 Nano flight controller.

Overview

The simulator implements a full 6-DOF rigid body model of a quadrotor with cascaded PID control, multi-sensor fusion, and a waypoint mission system.

Control

• Cascaded PID architecture — outer position loop generates desired attitude and thrust setpoints, passed to a high-rate inner attitude loop
• Gains derived analytically via pole placement on the linearised plant (attitude: ωn = 8 rad/s; position: ωn = 1.2 rad/s; 5-10x bandwidth separation)
• Derivative-on-measurement, integral anti-windup, and tilt compensation

State Estimation

• Kalman Filter (translational) — 6-state [x, y, z, vx, vy, vz] estimator fusing 100 Hz IMU predictions with 5 Hz GPS and 50 Hz altimeter corrections
• Complementary Filter (attitude) — fuses gyroscope integration with accelerometer tilt sensing, with linear acceleration compensation to prevent false tilt readings during maneuvers

Sensor Models

• Gyroscope: white noise + random-walk bias drift
• Accelerometer: body-frame specific force with noise
• GPS: 5 Hz position updates with configurable noise std dev
• Altimeter: 50 Hz with configurable noise std dev

Verified Performance

• Steady-state position error < 8 cm during waypoint hover
• 3-meter multi-axis step response settles in < 3 seconds
• Stable through full waypoint mission with sensor noise enabled

Roadmap

• Cascaded PID control with pole-placement gains
• Complementary filter attitude estimator
• Kalman filter position estimator
• Waypoint mission sequencing
• LQR attitude control
• LQR full-state control
• Minimum snap trajectory planning
• Extended Kalman Filter (EKF)
• C++ port with MAVSDK for CUAV V5 Nano
• Hardware-in-the-Loop (HIL) testing via Raspberry Pi companion computer

Structure

Project-Icarus/ ├── simulations/ │ ├── 1d_vertical_kinematics/ # 1D altitude testbed used to validate │ │ thrust model before extending to 6-DOF │ └── 6dof_full_flight/ # Full 6-DOF simulator │ ├── main_sim.py # Primary simulator (KF + CF sensor fusion) │ └── report/report.md # Detailed system architecture and results ├── fsw/ # Flight software (C++ / MAVSDK) — in progress ├── data_analysis/ └── requirements.txt

Getting Started

git clone https://github.com/brandonjacobson/Project-Icarus.git
cd Project-Icarus
pip install -r requirements.txt
python simulations/6dof_full_flight/main_sim.py

Dependencies

• numpy
• matplotlib
• scipy
• Pillow