Autonomous-Vehicle-Navigation (CSAI 301 Project)

This project consists of two phases:

1. Path Planning Algorithms Visualization and Comparison
2. Reinforcement Learning for Autonomous Navigation

---

📋 Phase 1: Path Planning Algorithms

Overview

An interactive system for visualizing and comparing different pathfinding algorithms through a GUI interface. The system allows users to select specific algorithms and observe their behavior in real-time on a grid-based environment.

---

🛠 Setup & Installation

1. Install Required Dependencies:

pip install -r requirements.txt

2. Run Files in the Following Order (if issues arise):

python libraries.py
python node_grid.py
python algorithms.py
python gui.py
python main.py

---

🔄 Running Options

Option 1: Algorithm Comparison (Backend)

Run algorithms.py to:

• Generate performance comparison of all algorithms.
• Create algorithms_results.zip containing summaries and visualizations.
• Display a comparison table in the terminal.

Option 2: Interactive GUI (Frontend)

Run main.py to:

• Launch the interactive GUI.
• Select specific algorithms to visualize.
• View real-time execution results.
• Compare selected algorithms' performance.

---

📂 Output Structure

• algorithm_outputs/: Main output directory containing:
  • Algorithm execution videos
  • Path visualization images
  • Exploration process visualizations
  • Performance statistics

---

🎯 Features

• Multiple algorithm implementations (BFS, DFS, UCS, IDS, Greedy, A*, Hill Climbing, Simulated Annealing).
• Real-time visualization of algorithm execution.
• Comparison of performance metrics (e.g., execution time, explored nodes).
• User-friendly interactive GUI interface.

---

---

📋 Phase 2: Reinforcement Learning Navigation

Overview

An implementation of Q-Learning-based Reinforcement Learning (RL) for autonomous navigation in a simulated grid environment. The agent learns to navigate efficiently while avoiding obstacles and reaching a goal state.

---

🛠 Setup & Installation

1. Install Required Dependencies:

pip install -r requirements.txt

2. Run the Simulation:

python main.py

3. Observe Real-Time Visualization:

• The agent's learning behavior will be displayed on the screen.
• Metrics will update live during each episode.

---

📊 Output

• learning_report.csv: A detailed report containing:

  • Episode Number
  • Steps Taken per Episode
  • Total Rewards Earned
  • Success/Failure Status

• Statistics Screen: After training, a final screen displays:

  • Total Episodes
  • Average Steps per Episode
  • Average Reward per Episode
  • Success Rate
  • Total Training Time

---

🎯 Features

• Real-time visualization of the learning process using Pygame.
• Adaptive learning parameters (alpha, gamma, epsilon).
• Post-training analytics through learning_report.csv.
• Visualization of agent behavior and policy convergence.

---

🏗 Project Structure

├── phase1/
│   ├── algorithms.py     # Path planning algorithms
│   ├── gui.py            # GUI implementation
│   ├── node_grid.py      # Grid system setup
│   ├── main.py           # Main GUI application
│   ├── libraries.py      # Required dependencies
│   ├── algorithm_outputs/ 
│       ├── videos/       # Algorithm execution videos
│       ├── images/       # Path visualization images
│       ├── stats/        # Performance statistics

├── phase2/
│   ├── main.py           # Main RL application
│   ├── agent.py          # Q-Learning Agent logic
│   ├── environment.py    # Environment setup
│   ├── visualization.py  # Real-time visualization
│   ├── learning_report.csv  # Training metrics output

├── requirements.txt      # Required dependencies
├── README.md             # Project documentation

---

🔍 Requirements

• Python 3.12
• NumPy
• Matplotlib
• OpenCV
• Tkinter
• Imageio
• Pygame

---

📝 Notes

• Phase 1: Focused on comparing and visualizing pathfinding algorithms with static search strategies.
• Phase 2: Explored dynamic learning behavior through Q-Learning in a grid environment.
• Performance analytics and visualizations are provided for both phases.