Plane Localization with Particle Filter

Overview

A simple 1D/2D simulation of a plane localizing itself using a downward distance sensor and particle filtering.

How It Works

• Setup: Plane moves forward/backward; sensor measures altitude to ground.

• Goal: Estimate the plane's position using a predefined ground map.

• Method:

  1. Start with multiple guesses (particles).
  2. Move slightly, measure distance.
  3. Update and resample particles to keep only plausible guesses.
  4. 10% of particles remain random to help recovery.

Inspired by: Particle Filter Localization - YouTube

Limitations

• Requires a predefined map.
• Primarily 1D (optionally 2D with altitude changes).
• Uses a single downward ray.
• Assumes uniform motion (handcrafted odometry).
• Convergence can be slow.