{"id":29613790,"url":"https://github.com/barathkumarpm/multi_agent_lunar_landing_sim","last_synced_at":"2025-07-20T22:37:17.683Z","repository":{"id":304953244,"uuid":"1020677362","full_name":"BarathKumarpm/Multi_Agent_Lunar_Landing_Sim","owner":"BarathKumarpm","description":null,"archived":false,"fork":false,"pushed_at":"2025-07-16T08:26:53.000Z","size":125,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-07-17T11:44:39.487Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/BarathKumarpm.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null}},"created_at":"2025-07-16T08:23:15.000Z","updated_at":"2025-07-16T08:26:56.000Z","dependencies_parsed_at":"2025-07-17T16:02:24.529Z","dependency_job_id":null,"html_url":"https://github.com/BarathKumarpm/Multi_Agent_Lunar_Landing_Sim","commit_stats":null,"previous_names":["barathkumarpm/multi_agent_lunar_landing_sim"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/BarathKumarpm/Multi_Agent_Lunar_Landing_Sim","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/BarathKumarpm%2FMulti_Agent_Lunar_Landing_Sim","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/BarathKumarpm%2FMulti_Agent_Lunar_Landing_Sim/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/BarathKumarpm%2FMulti_Agent_Lunar_Landing_Sim/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/BarathKumarpm%2FMulti_Agent_Lunar_Landing_Sim/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/BarathKumarpm","download_url":"https://codeload.github.com/BarathKumarpm/Multi_Agent_Lunar_Landing_Sim/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/BarathKumarpm%2FMulti_Agent_Lunar_Landing_Sim/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":266210974,"owners_count":23893346,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":[],"created_at":"2025-07-20T22:37:14.625Z","updated_at":"2025-07-20T22:37:17.677Z","avatar_url":"https://github.com/BarathKumarpm.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 🛰️ Multi-Agent Lunar Lander Simulation using Reinforcement Learning\n\nA simulation environment inspired by OpenAI Gym and PettingZoo that explores autonomous coordination between **two lunar landers** attempting simultaneous touchdowns. This project focuses on extending the traditional single-agent Lunar Lander task to a **multi-agent reinforcement learning (MARL)** scenario — introducing complexity, real-time cooperation, collision avoidance, and fuel optimization.\n\n---\n\n## 📘 1. Introduction\n\n**Multi-Agent Lunar Lander** is an advanced version of the classical Lunar Lander environment built using **OpenAI Gymnasium (Box2D)** and the **PettingZoo** framework. Instead of controlling a single lander, this project introduces two landers that must coordinate to land safely — transforming the task from a single-agent control problem into a **multi-agent coordination challenge**.\n\nAgents must manage:\n- Independent and interdependent controls\n- Thrust, orientation, fuel optimization\n- Inter-agent interference and stability management\n\nThe simulation resembles **real-world scenarios** such as:\n- Coordinated landings of autonomous spacecraft\n- Drone fleet coordination\n- Multi-robot system control in unstructured environments\n\n---\n\n## 🚩 1.2 Problem Statement\n\nIn future space missions, simultaneous landings of multiple landers on the Moon could help reduce costs via:\n- Shared payload capacity\n- Rideshare opportunities\n- Reusability\n\nHowever, this approach increases:\n- Navigational complexity\n- Development cost for hazard avoidance and control systems\n- Risk of interference and failure\n\nThis simulation environment allows for experimentation and training of agents under such constraints, enabling researchers to optimize for cost, safety, and performance.\n\n---\n\n## 🎯 1.3 Objectives\n\nThe project’s objectives include:\n- Simulating safe and fuel-efficient landings in a multi-agent scenario\n- Designing adaptive control strategies using reinforcement learning\n- Building agents that generalize across dynamic conditions and unforeseen environmental states\n- Benchmarking various RL algorithms using performance metrics such as:\n  - Landing success\n  - Fuel efficiency\n  - Collision avoidance\n  - Time-to-land\n\n---\n\n## ✅ 1.4 Benefits of Simultaneous Multi-Lander Missions\n\n- **Cost Efficiency**: Reduces need for redundant backup systems\n- **Mission Reliability**: Promotes robust risk assessment and coordination\n- **Improved Precision**: Helps test real-time autonomous landing under limited zone constraints\n- **Scalability**: Encourages multi-lander, multi-mission automation\n\n\u003e Synchronization between landers is crucial — failure in coordination can increase mission time and operational costs.\n\n---\n\n## 🔭 1.5 Scope of the Project\n\nThis project:\n- Converts OpenAI’s single-agent Lunar Lander into a **multi-agent PettingZoo-compatible environment**\n- Simulates real-world lunar dynamics such as:\n  - Irregular terrain\n  - Varied soil types\n  - Light and gravity conditions\n\n### Key Features:\n- **Dual-agent control** with independent and shared policy learning\n- **Custom reward functions** for balancing:\n  - Safe landing\n  - Fuel use\n  - Synchronization\n  - Collision avoidance\n- **Parameter tuning** for different operation scenarios\n- **Modular design** for research extensibility and reproducibility\n\n---\n\n## 🛠️ Technologies \u0026 Tools\n\n- `OpenAI Gym`\n- `PettingZoo`\n- `Box2D`\n- `Stable-Baselines3`\n- `Python 3.x`\n- `NumPy`, `Matplotlib`, etc.\n\n---\n\n## 📈 Real-World Inspiration\n\nInspired by **NASA's CADRE and AAMAS projects**, which explore decentralized coordination among autonomous robotic landers.\n\n---\n\n## 🤝 Contributions \u0026 Future Work\n\nWe welcome contributions! Potential extensions:\n- Inter-agent communication modeling\n- Competitive vs cooperative multi-agent settings\n- Integration with real-world sensor data\n\n---\n\n## 📄 License\n\nThis project is open-sourced under the MIT License.\n\n---\n\n## 🌌 Final Note\n\nMulti-Agent Lunar Lander provides a challenging yet promising platform for advancing reinforcement learning in space robotics, autonomous control, and multi-agent systems. Through research and collaboration, this project aims to serve as a testbed for future intelligent space missions.\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbarathkumarpm%2Fmulti_agent_lunar_landing_sim","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbarathkumarpm%2Fmulti_agent_lunar_landing_sim","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbarathkumarpm%2Fmulti_agent_lunar_landing_sim/lists"}