Project Overview

Designed and controlled a nonlinear, coupled MIMO aerodynamic system from first principles, validated in simulation and on real hardware.

• System type: Nonlinear, coupled, underactuated
• DOF: 2 (Pitch & Yaw)
• Tools: MATLAB/Simulink, Kalman Filter, SolidWorks, Embedded C
• Focus: Modeling, estimation, control

This project involved the complete design of a Twin Rotor Aerodynamic System (TRAS), a benchmark platform for helicopter-like dynamics and multivariable control research. The system exhibits strong nonlinearities and aerodynamic coupling, making it suitable for advanced control and state estimation studies.

The Twin Rotor Aerodynamic System (TRAS) is a laboratory-scale multi-input multi-output (MIMO) system that emulates the dynamics of a helicopter through the interaction of aerodynamic forces, gravitational effects, and strong axis coupling. The objective of this project is to develop a physics-based dynamic model of the TRAS and design a feedback control strategy capable of stabilizing and accurately tracking reference commands for both pitch and yaw angles.

To provide a clear understanding of the system structure and signal flow, the overall control architecture of the TRAS is illustrated below. This block diagram highlights the relationship between control inputs, system dynamics, sensor feedback, and the closed-loop controller, serving as a roadmap for the modeling and control design steps presented in the following sections.

System Architecture

The TRAS consists of two independently actuated rotors mounted on a pivoted structure, providing pitch and yaw motion. Due to aerodynamic interaction, gravitational effects, and actuator dynamics, the system behaves as a strongly coupled MIMO plant.

Mechanical Design & CAD

The mechanical structure was designed to ensure adequate stiffness, controlled mass distribution, and predictable dynamic behavior. Special attention was given to center-of-gravity placement and inertia properties to improve controllability and reduce undesired coupling.

Mathematical Modeling

The system was modeled as a nonlinear, coupled dynamical system using a Newton–Euler formulation. The equations of motion capture gravitational effects, aerodynamic thrust, drag, and cross-axis coupling between pitch and yaw.

The resulting model exhibits:

• Nonlinear input–output behavior
• Strong coupling between states
• Sensitivity to parameter uncertainty

Nonlinear state-space model of the Twin Rotor Aerodynamic System

Simulink Implementation

A high-fidelity Simulink model was developed directly from the derived equations. The model was structured in a modular form separating plant dynamics, actuators, sensors, and estimation logic, enabling rapid iteration and controller testing. The TRAS exhibits highly nonlinear dynamics due to aerodynamic coupling between the rotors and gravitational effects. A nonlinear mathematical model was developed to capture these interactions, using experimentally identified parameters and physical laws governing rotor thrust and moments. The model structure shown below outlines the relationship between actuator inputs, aerodynamic forces, and the resulting angular motions of the system.

Full Simulink overview

EKF implementation in Simulink with Realtime data input from the model

Control Strategy

Control laws were designed using a control-oriented model derived from the nonlinear dynamics. Although pitch and yaw were actuated independently, the controller explicitly accounts for cross-axis coupling. Estimated states from the Kalman Filter were used for feedback.

Embedded & Real-Time Implementation

The control and estimation algorithms were deployed on a microcontroller for real-time execution. The embedded architecture mirrors the simulation model, ensuring consistency between simulated and experimental behavior.

Given below is the link to the code implemented on the microcontroller.
Note: An OLED-based interactive menu system was initially implemented for on-device tuning and diagnostics; it was later removed from the portfolio version to highlight the real-time sensing, control, and actuation architecture.

Key Engineering Takeaways

• Nonlinear modeling of aerodynamic systems
• Control of coupled MIMO plants
• Practical state estimation using Kalman filtering
• Real-time implementation challenges