Skip to content

README.md

Latest commit

 

History

History
64 lines (45 loc) · 3.47 KB

README.md

File metadata and controls

64 lines (45 loc) · 3.47 KB

CORODRO Algorithm Sandbox

Teaching-oriented archive of prototype scripts, robotics experiments, and algorithm exercises that informed the CORODRO project. This repository is the right place to learn ideas quickly without loading the full mission stack.

The filenames in this repo have been normalized to descriptive snake_case names so students can understand the intent of a script before opening it.

Most scripts now parse under Python 3. The top-level perception and teleoperation demos still require a ROS 1 environment with packages such as rospy, sensor_msgs, geometry_msgs, and cv_bridge.

Added Value

  • Small scripts that are easier to read than the mission code.
  • Early robotics experiments preserved as learning material.
  • A student-friendly catalogue of planning, localization, and reinforcement-learning examples.

Algorithms Used

  • Occupancy-grid mapping from lidar data.
  • ROS image and OpenCV bridging.
  • Q-learning and simple MDP examples.
  • A*, D*, FastSLAM, EKF, MPC, Stanley control, and quintic polynomial planning examples.
  • Experimental DWA plus A* integration.

Repository Map

  • ai_general/: reinforcement learning and optimization exercises.
  • dwa_recursive/: experimental DWA implementation with A* support.
  • legacy_navigation_planning/: older planning and SLAM examples not used directly in the final mission stack.
  • top-level scripts: small ROS mapping, localization, teleoperation, and perception demos.

Quick Start

  1. Start with ai_general/ if you want standalone Python examples with no ROS dependency.
  2. Move to legacy_navigation_planning/ for readable reference implementations of classical robotics algorithms.
  3. Use the top-level scripts only once you already have ROS 1 and the sensor/message dependencies installed.
  4. Treat dwa_recursive/ as an experimental sandbox rather than a polished planner package.

Recommended Learning Path

  1. Read the top-level ROS examples if you are new to ROS nodes.
  2. Move to ai_general/ for simple algorithm exercises.
  3. Explore legacy_navigation_planning/ for classic robotics algorithms.
  4. Finish with dwa_recursive/ if you want to study a more integrated planning experiment.

Top-Level Script Guide

  • depth_camera_cv_bridge_demo.py: ROS/OpenCV demo that converts depth-camera data into a colorized DEM-style visualization.
  • fake_localization_node.py: ROS helper that publishes simple localization information for navigation tests.
  • fake_mapper_node.py: ROS helper that publishes a lightweight fake map for planner testing.
  • keyboard_teleop_experiment.py: ROS keyboard teleoperation experiment.
  • lidar_occupancy_map_builder.py: ROS lidar-to-occupancy-map prototype.
  • lidar_scan_to_occupancy_grid.py: ROS LaserScan-to-OccupancyGrid utility.
  • ros_image_to_opencv_bridge.py: reusable ROS image bridge for OpenCV processing experiments.

Best First Files

  • ai_general/simple_q_learning_demo.py: smallest reinforcement-learning example in the repo.
  • ai_general/dynamic_programming_knapsack.py: compact dynamic-programming example.
  • legacy_navigation_planning/quintic_polynomial_planner.py: clean trajectory-generation example.
  • dwa_recursive/a_star_planner.py: readable grid-search helper before opening the larger DWA experiment.

Notes For Students

  • This repo is intentionally eclectic. Think of it as a lab notebook that has been cleaned up into a teaching archive.
  • The code quality varies by file because some scripts were written as prototypes or learning exercises.