2D Rocket Simulation

A real-time 2D self-landing rocket simulation with full-state feedback LQR control, using SFML for visualization, Eigen for linear algebra, and matplotlib for data plots. The simulation demonstrates how different LQR gain profiles affect rocket landing behavior.

Overview

• Physics: Simulates a 2D rigid-body rocket with 6-DOF (x, y, θ and their derivatives), gravity, thrust, and a variable-thrust gimbaled engine.
• Control: Uses a Linear Quadratic Regulator (LQR) for full-state feedback to stabilize the rocket at a desired state equilibrium point. Two profiles are demonstrated:
  • Slow: A slower, more gentle response profile that prioritizes minimizing control actuation effort, characterized by a high control cost.
  • Fast: A faster, more aggressive response that stabilizes the rocket more rapidly at the expense of greater control effort.
• Visualization: Real-time graphics with SFML, displaying rocket state, control actions, and system behavior over time.

How to run the simulation:

1. LQR Gain Generation:
   Run the Python script to generate LQR gain matrices for both profiles:

   python scripts/compute_lqr_gains.py
   

   This produces lqr_gains_slow.json and lqr_gains_fast.json.

2. Build the Project:

   • Prerequisites: CMake 3.16+, GCC 14+, SFML 3.0+, Eigen 5.0+.
   • Configure and build:

     cd build
     cmake ..
     cmake --build .
     

3. Run the Simulation:
   From the build directory:

   ./RocketSim.exe slow   # Gentle controller
   ./RocketSim.exe fast   # Aggressive controller
   

   The controller profile determines which gain file is loaded.