Void-Navigator: A Real-Time Space Path Planning System Based on the A* Algorithm

Void-Navigator is an innovative space navigation simulator that combines the classic A* path planning algorithm with real-time near-Earth object data from NASA. Its core goal is to enable a virtual spacecraft to safely navigate from Earth to Mars through a dynamic environment of real asteroids. Developed as part of the Spring 261 CSE 3812 (E) - Artificial Intelligence Lab course, this project demonstrates the practical application of AI algorithms in solving real-world path planning challenges.

Project Details:

• Author/Maintainer: rahad404
• Source: GitHub (link: https://github.com/rahad404/Void-Navigator)
• Release Date: 2026-06-13

Core Purpose: The project creates an immersive environment where users guide a virtual spacecraft through a 2D grid filled with obstacles derived from NASA's real asteroid observations. Each run presents a unique challenge due to the dynamic nature of the data.

Key Feature: Direct integration with NASA's NeoWS API to fetch real-time asteroid data as navigation obstacles.

A Algorithm Implementation*:

• Node Class: Models grid cells with g (actual cost), h (heuristic cost), f (total cost = g+h), parent reference, and hazard weight.
• Heuristic Function: Defaults to Octile distance (optimal for 8-direction movement).
• Priority Queue: Uses Python's heapq for efficient node selection.

Grid System:

• 40x40 grid with 8-direction movement (orthogonal cost:1.0, diagonal cost:~1.414).
• Supports dynamic obstacle management and grid reset.

Hazard Weight Mechanism:

• Creates progressive danger zones around obstacles using distance decay (weight = cost_increment/distance).
• Simulates gravity traps and debris fields; cost calculation includes hazard weight.

Cost Formula: tentative_g = current.g + step_cost + neighbor.hazard_weight

NeoWS API Usage:

• Endpoint: https://api.nasa.gov/neo/rest/v1/feed
• Data Fetch Flow: Construct URL → send HTTP request (5s timeout) → parse JSON → map to grid coordinates.

Data Mapping:

• Extracts: ID, name, diameter (avg of min/max), velocity, hazard status.
• Coordinate Generation: Uses MD5 hash of asteroid ID for deterministic, uniform grid placement.

Offline Backup:

• Includes 16 famous asteroids (e.g., Apophis, Bennu, Ceres) for network failure scenarios.

Stellar Data:

• Famous stars: Sirius (brightest), Vega (summer iconic), Betelgeuse (red supergiant), Polaris (navigation reference).
• Each entry includes spectral type, apparent magnitude, and distance.

Deep Space Objects:

• Galaxies: Andromeda (nearest large galaxy).
• Nebulae: Orion Nebula (star formation), Crab Nebula (supernova remnant).

These elements enhance the visual immersion of the simulation.

Coordinate Strategy:

• MD5 hash ensures same asteroid ID maps to same grid position.
• Avoids boundary areas (margin=3) and start/end zones (Earth at 5,5; Mars at grid_size-6, grid_size-6).

Error Handling:

• API failure → graceful fallback to offline data.
• Network timeout (5s) to prevent UI freeze.
• Clear error messages for missing environment variables (e.g., NASA API key).

Educational Concepts:

• Search algorithms (A* implementation and heuristic design).
• Data integration (API calls and exception handling).
• Visualization preparation (grid system for Pygame).
• Physical simulation (hazard weight for continuous space effects).

Future Extensions:

• Add dynamic obstacles (moving asteroids).
• Support multi-spacecraft collaborative planning.
• Introduce fuel constraints or other optimization goals.
• Include more celestial types (comets, satellites).

Void-Navigator successfully merges classic AI algorithms with real-world astronomical data to create an educational and entertaining space navigation simulator. It demonstrates the practicality of A* in dynamic environments and showcases robust engineering practices (modular design, error handling). The project bridges theoretical AI concepts with real data, making it a valuable learning tool for AI students.