RoboCleaner: A Q-Learning Robot Simulation

RoboCleaner is a C++ application that simulates a robotic cleaner learning to navigate and clean a room using Reinforcement Learning (Q-Learning). The robot learns to avoid walls and maximize cleaning efficiency over time.

Features

• Q-Learning Agent: Implements a tabular Q-learning algorithm to learn optimal policies.
• Interactive UI: Visualize the cleaning process, room layout, and real-time statistics (Clean %, Moves, Exploration Rate).
• Customizable Environment: Adjustable obstacle density and simulation speed.
• Real-time Graphs: Monitor the agent's learning progress with ImGui-based charting.

Prerequisites

• CMake (3.14 or higher)
• C++ Compiler (supporting C++17)

Building the Project

1. Clone the repository:

   git clone <repository_url>
   cd RoboCleaner./

2. Create a build directory and compile:

   mkdir -p build
   cd build
   cmake ..
   make

3. Run the application:

   ./RoboCleaner

Controls

• Start/Stop: Toggle the simulation.
• Reset: Reset the environment and the agent's memory (Q-Table) to start fresh.
• Speed: Adjust the simulation speed (train faster or watch closely).
• Obstacle Density: Change the number of walls in the generated map (automatically resets).

Technologies Used

• C++17
• Dear ImGui for UI
• GLFW for window management
• OpenGL for rendering