Project brief

Evaluating autonomous flight control loops, trajectory tracking, and wide-area navigation algorithms requires highly accurate simulation models that mirror the physical aircraft. However, virtual models often rely on generalized approximations, completely ignoring critical real-world parameters like the true asymmetric 3x3 inertia tensor, local center of mass shifts, current wiring restrictions, and electrical power system dynamics. This project investigates the systematic reverse-engineering and multi-platform deployment of an operational micro-quadcopter equipped with a Holybro Kakute H7 V2 Flight Stack running PX4 Autopilot. By constructing an unconstrained, 6-Degrees of Freedom (6-DoF) "Digital Twin," this work aims to create a database-driven system description. The unified dataset ensures cross-platform symmetry, allowing identical plant parameters to drive predictive physics simulations in Gazebo and, model-based control loop shaping in MATLAB/Simulink.

Intended project outputs

• Develop an Unconstrained Master CAD Assembly: Recreate the standalone physical quadcopter frame and compile a fully populated standard parts library for all propulsion and structural fasteners.
• Model the Electronics Stack and Space Constraints: Digitally reconstruct the Kakute H7 V2 flight controller and Tekko32 4-in-1 ESC stack, mapping exact physical clearances, mounting patterns, and high-current wiring paths.
• Establish a Software-Agnostic Parameter Database: Centralize all derived physical, inertial, and electrical parameters into a unified format (.json/.csv) to serve as a single source of truth for downstream solvers.
• Deploy a Parallel Simulation Framework: Generate compliant robot description trees (URDF/SDF) optimized for Gazebo physics alongside automated structural models inside MATLAB Simscape Multibody.
• Characterise Electro-Mechanical and Battery Dynamics: Build equivalent mathematical circuit models capturing the non-linear discharge, voltage sag, and internal resistance of a LiPo battery coupled with Brushless DC motor torque constants.
• Map Cyber-Physical PX4 Configurations: Extract geometric motor-arm coordinates and IMU reference axes from the digital twin to directly parameterize PX4 control allocation and sensor rotation tables.
• Incorporate Flight Readiness Hardware: Design and integrate modular, lightweight 3D-printed landing stabilizers and secure battery strapping mounts into the final airframe layout.

Skills you will develop

• CAD Design and System Metrology
• Computational Rigid-Body Dynamics (Center of mass tracking, 3x3 inertia matrix extraction, 6-DoF free-flight kinematic mapping)
• Cyber-Physical System Simulation (Gazebo SDF/URDF structure, MATLAB Simscape, and Simulink model-based design)
• Electrical Plant Parameterization (Equivalent circuit battery modeling, actuator profiling, and PX4 Autopilot configuration)