RL Based Inverted Pendulum over Wireless Channel

Wired + Wireless Implementation

---

Overview

This project implements control of an inverted pendulum system using both wired (Arduino-based) and wireless (ESP-NOW-based) architectures.

The system is modeled and analyzed in MATLAB/Simulink, and control is implemented using a proportional controller derived via pole placement techniques.

---

System Architecture

1. Wired Control (Arduino)

• Sensor readings are directly obtained by the Arduino
• Control computation is performed locally
• Output voltage is applied directly to the motor

2. Wireless Control (ESP32 + Arduino)

• Sensor data is collected by an ESP32 (plant side)
• Data is transmitted via ESP-NOW
• Controller ESP32 computes control input
• Control signal is sent back to plant ESP
• Arduino Mega acts as a voltage interface (3.3V → 5V) for motor actuation

---

Repository Structure

proportional/
    Proportional_controller.ino

simulation/
    controlled_simulation.m
    linearization_with_motor.m
    pendulum_model.slx
    pole_placement_with_motor.m

wireless/
    Delay_RTT.ino
    ESP_PWM_reader.ino
    Pendulum_controller.ino
    Pendulum_plant.ino

---

Simulation (MATLAB / Simulink)

Files

File	Description
linearization_with_motor.m	Linearizes the pendulum system including motor dynamics
pole_placement_with_motor.m	Computes controller gains using pole placement
controlled_simulation.m	Simulates closed-loop system performance
pendulum_model.slx	Simscape model of pendulum + motor dynamics

Purpose

• Model system dynamics
• Include motor voltage behavior
• Design controller gains before hardware implementation

---

Proportional Control (Wired Implementation)

File: proportional/Proportional_controller.ino

• Implements proportional controller on Arduino
• Uses gains obtained from simulation
• Sensor readings are directly processed
• Motor is controlled via wired interface

---

Wireless Control (ESP-NOW Implementation)

Pendulum_plant.ino

• Runs on ESP32 (plant side)
• Reads sensor data
• Transmits data to controller via ESP-NOW
• Receives control signal

Pendulum_controller.ino

• Runs on ESP32 (controller side)
• Receives sensor data
• Computes control input (voltage)
• Sends control signal back to plant

ESP_PWM_reader.ino

• Runs on Arduino Mega
• Reads PWM signal from ESP32 (3.3V)
• Converts it to 5V signal for motor driver
• Acts as a level shifter + actuator interface

Delay_RTT.ino

• Measures communication delay (Round Trip Time)
• Evaluates latency in wireless control loop

---

Key Features

• Dual architecture: wired and wireless control
• Real-time ESP-NOW communication
• Latency measurement and analysis
• MATLAB-based system modeling and controller design
• Integration of embedded systems with control theory

---

Applications

• Networked control systems
• Wireless feedback control
• Real-time embedded control systems
• Robotics and automation

---

Notes

• Controller gains are derived from MATLAB simulations
• Wireless system introduces delay, analyzed via RTT measurement
• Arduino Mega is used to handle voltage level mismatch (3.3V → 5V)