2D Graph-Based SLAM

This repo is no longer maintained.

Pose Graph SLAM has taken its place.

A Python implementation of 2D graph-based SLAM using:

• Point-to-Point ICP for scan matching
• Pose graph optimization
• Loop closure detection
• Occupancy grid mapping

Features

• Real-time 2D lidar SLAM
• Pose graph with loop closure
• Occupancy grid generation for path planning
• Visualization tools

Files

• slam.py - Main SLAM implementation
• Scan_Pose_Data/ - Sample datasets

Requirements

• numpy
• scipy
• matplotlib

Output

Add files to post-process raw OGM from SLAM

• Tweak angle to square up alignment
• Add inflation

slam_dev/
├── slam.py                   # SLAM code (includes OGM class)
├── occupancy_grid.py         # New standalone OGM class (with added methods)
├── maps/
│   ├── raw_map.npz           # From SLAM
│   ├── raw_map.png           # Visualization
│   ├── aligned_map.npz       # After rotation
│   ├── final_map.npz         # After rotation + inflation
│   └── final_map.png         # Final visualization
├── diagnostic.py             # Script to examine contents of OGM cells
├── align_auto_pca.py         # Script to process raw map
├── align_manual.py           # Script to process raw map
└── path_planner.py           # Script for path planning