Thrust Vector Control

Thrust vectored inverted pendulum stabilized via PID controller (Video).

Authors

• Jonathan Cochran
• Bryson Jaipean
• Cameron Retzlaff

Background

• Thrust vector control is often used to steer an aircraft or rocket by adjusting the direction of thrust
• This project forms a basis for a control model for a thrust vectored rocket

Goals

• Adjust angle of thrust to keep inverted pendulum upright
• Gather initial controller feedback for a thrust vectored rocket
• Evaluate performance of PID control system

System Design

• System consumes rotation angle with an analog input from a gyroscope
• Positional servo adjusts angle of thrust, manipulating the torque about the point of rotation
• Thrust force generated via Brushless RC Motor connected to a RC Plane Propeller
  • Motor throttled by electronic speed controller with a wireless RC transmitter
• PID control and feedback powered by Arduino Nano
  • Enabled constant adjustment to thrust force angle
• Parts mounted to a wooden dowel
• Inverted pendulum affect created by ensuring the end opposite the propeller weighted higher than propeller end
• Moment of inertia calculated from distance between center of mass and rotation point
• Breadboard position adjusted to create multiple test cases of differing center of mass

Pendulum Balancing

Free-body Diagram

Torque applied about point of rotation from thrust force F at angle α, modifying pendulum rotation angle θ

Wiring Diagram

PID Controller

• A PID (Proportional-Integral-Derivative) controller is a feedback control loop
• Enables constant modulated control of a system

Stable System

• Tested for ultimate proportional gain and ultimate period of a regular pendulum
• Determined Zeigler Nichols PID values for naturally stable system
• Implemented PID values to control stable system
• PID system returned 5 second stabilization improvement over natural system at 110 degree displacement

Natural System vs. PID Controlled System

Inverted System

• Ziegler-Nichols method PID values not sufficient for inverted system
• Evaluated divergence time to determine to validate Pessen Integral PID values
• Tuned PID values by experimentally adjusting Pessen Integral PID values
• Implemented +/- 2 degree deadband at center to avoid PID accumulating error
• Evaluated system control with two center of mass locations
  • Location 1: 4in from center of rotation
  • Location 2: 4.5in from center of rotation

Stable System Configuration 1

Stable System Configuration 2

Key learnings

• Proportional value proportionally connected to the moment of inertia of the system
• Integral value is inversely proportional to moment of inertia of the system
• Derivative term not impacted by the moment of inertia of the system