🤖 AI Chase Tag Game

An interactive web-based game where two AI agents compete in a pursuit-evasion scenario. Watch as the chaser uses A* pathfinding to hunt down the evader, while the evader employs sophisticated behavioral AI to escape and survive.

🎮 Live Demo

https://notsogenius69.github.io/Chase-Tag-Game/

📖 Overview

This project explores the fascinating world of adversarial AI in game development. Two AI agents take turns being the chaser and evader in 20-second rounds, demonstrating different algorithmic approaches to pursuit and evasion.

Key Features

• Intelligent Chaser AI: Uses A* pathfinding algorithm for optimal pursuit
• Strategic Evader AI: Employs behavioral AI with:
  • Real-time threat assessment
  • Dead-end and corner detection
  • Escape route analysis
  • Strategic positioning
  • Wall-following behavior
• Dynamic Gameplay: Automatic role switching after each round
• Obstacle Navigation: Both agents navigate around walls and obstacles
• Score Tracking: Keeps track of wins for both players
• Pure Vanilla JS: No frameworks or dependencies required

🎯 How It Works

Chaser Strategy (A* Pathfinding)

The chaser uses the A* algorithm to find the optimal path to the evader:

• Calculates shortest path considering obstacles
• Updates path in real-time as evader moves
• Efficient grid-based pathfinding

Evader Strategy (Behavioral AI)

The evader uses a sophisticated multi-layered behavioral system:

1. Situation Analysis: Evaluates current danger level and available escape routes

2. Position Safety Evaluation: Checks for dead ends, corners, and open space

3. Behavioral Selection:

   • Emergency Dodge: When chaser is very close (< 60px)
   • Dead End Escape: When trapped with ≤2 escape routes
   • Strategic Positioning: When safe, moves to positions with maximum escape options
   • Smart Evasion: Normal mode with predictive movement

4. Scoring System: Evaluates potential moves based on:

   • Number of escape routes
   • Distance from boundaries
   • Open space availability
   • Distance from chaser
   • Trap avoidance

🚀 Getting Started

Prerequisites

• A modern web browser (Chrome, Firefox, Safari, Edge)
• No additional dependencies required!

Installation

1. Clone the repository:

git clone https://github.com/yourusername/ai-chase-tag-game.git

2. Navigate to the project directory:

cd ai-chase-tag-game

3. Open index.html in your web browser:

# On macOS
open index.html

# On Linux
xdg-open index.html

# On Windows
start index.html

Or simply drag and drop index.html into your browser.

🎲 Game Rules

1. Objective:

   • Chaser: Tag the evader before time runs out
   • Evader: Avoid being tagged for 20 seconds

2. Scoring:

   • Chaser wins if they tag the evader
   • Evader wins if they survive 20 seconds

3. Role Switching: After each round, players switch roles automatically

4. Victory: First to 5 wins (game ends automatically)

🛠️ Technical Implementation

Technologies Used

• HTML5 Canvas: For rendering the game
• Vanilla JavaScript: Game logic and AI algorithms
• CSS3: Styling and UI design

Core Algorithms

A* Pathfinding (Chaser)

// Simplified A* implementation
f(n) = g(n) + h(n)
- g(n): Cost from start to node n
- h(n): Heuristic (Euclidean distance to goal)

Behavioral AI (Evader)

// Multi-layered decision system
1. Analyze current situation
2. Evaluate position safety
3. Generate movement candidates
4. Score each candidate
5. Execute best move

Project Structure

ai-chase-tag-game/
│
├── index.html          # Main HTML file with embedded CSS and JavaScript
├── README.md           # This file
├── screenshot.png      # Game screenshot (add your own)
└── LICENSE             # License file (optional)

📊 AI Learning Journey

Initial Approach (Failed ❌)

• Used A* for both chaser and evader
• Problem: Evader kept getting trapped in corners
• Why: A* is designed for single-agent shortest path, not adversarial scenarios

Solution (Success ✅)

• Kept A* for chaser (works great for pursuit)
• Implemented behavioral AI for evader with:
  • Spatial awareness
  • Trap detection
  • Strategic decision making
  • Real-time adaptation

Key Insight

"A* finds the shortest path to a goal. But in adversarial games, the 'goal' is dynamic and the opponent is actively trying to counter you. Single-agent algorithms don't translate directly to multi-agent scenarios."

⭐ If you found this project interesting, please consider giving it a star!