This repository contains a MATLAB simulation of a quadcopter with different control strategies and dynamics.

Simply execute either script directly in MATLAB - run attitude_control.m for basic attitude stabilization or position_control.m for more comprehensive position control scenarios with customizable parameters.

• Nonlinear quadcopter dynamics
• Motor dynamics (1st order)
• Sensor noise
• Wind disturbances
• Switchable controller: Choice of PID, SMC, or INDI (set controller_type at the top of the script)
• Slung load dynamics

quadcopter-control/
├── attitude_control.m 
├── position_control.m         # Main simulation scripts
├── functions/
│   ├── slung_load_dynamics.m  # Slung dynamics calculation
│   ├── slung_compensator.m    # Swing damping
│   ├── rotor_thrust_mapping.m # Thrust allocation matrix
│   ├── helix_trajectory.m     # Helical path generator
│   ├── lpf.m                  # Low pass filter
│   ├── sat.m                  # Saturation function
│   └── log_and_plot.m         # Results visualization
├── logs/                      # Logs are automatically recorded here
├── figures/                   # Figures are automatically recorded here

• The quadrotor is modeled as a rigid body

• 1st order motor dynamics are included

• Sensor noises are incorporated in the model

• Wind disturbances are applied as both vertical forces and torques

• The cable has constant length and is massless

• The load is modeled as a point mass

• Movement in the x–z and y–z planes is independent (decoupled dynamics)

The translational accelerations are computed as:

ax = ( (-cos(phi)*sin(theta)*cos(psi) - sin(phi)*sin(psi)) * Fz ...
      + Fx_slung + Fx_wind ) / m_total;

ay = ( (-cos(phi)*sin(theta)*sin(psi) + sin(phi)*cos(psi)) * Fz ...
      + Fy_slung + Fy_wind ) / m_total;

az = ( cos(phi)*cos(theta)*Fz + Fz_slung + Fz_wind ) / m_total - g;

The torques about each axis are computed from the differences in rotor thrusts, arm length, and yaw drag coefficient. Angular accelerations are then:

dp = tau_phi / Ixx

(similar for dq, dr)

Each motor's thrust is modeled as a first-order lag:

T_dot = (T_command - T_actual) / tau_motor

This simulates the delay between commanded and actual thrust.

Sensor measurements for altitude, velocity, angles, and angular rates are corrupted with Gaussian noise to simulate real-world sensor imperfections.

ddalpha = - (g / L) * sin(alpha) ...  % gravity effect
          - d * dalpha ...            % damping
          - ax_est / L;               % quad acceleration

% Rotational acceleration in y axis
% Normally coupled but due to singularity it was assumed as decoupled
ddbeta = - (g / L) * sin(beta) ...    
         - d * dbeta ...           
         - ay_est / L;