EOS-SLAM

Event-Optimized Stereo-Inertial-LiDAR SLAM

EOS-SLAM is a real-time hybrid SLAM system that fuses stereo cameras, IMU, LiDAR, and wheel odometry into a single, efficient architecture. Designed with a fixed-size fusion window and adaptive triggering, it delivers consistent O(1) computational performance per frame while maintaining high accuracy and robust mapping.

---

🚀 Features

• ✅ Real-time stereo + IMU + LiDAR fusion
• ✅ O(1) computational cost per frame
• ✅ Event-driven updates (e.g., pose delta, info gain)
• ✅ Hybrid 2D occupancy + 3D voxel map
• ✅ ROS 2 ready & modular C++ codebase
• ✅ Benchmarking support (KITTI, EuRoC, TUM)

---

📁 Repository Structure

• fusion_core/: Sensor processing & fusion logic
• mapping/: 2D log-odds map + 3D local voxel maps
• localization/: EKF / trigger manager
• semantic/: Optional MobileNetV3-based semantic layer
• tools/: Benchmark logger + RViz plugins
• docs/: System design & benchmarking results

---

📦 Datasets & Evaluation

EOS-SLAM is tested on:

• KITTI (Stereo + IMU + GPS)
• EuRoC MAV (Stereo + IMU)
• TUM RGB-D (for stereo fallback testing)

Evaluation is performed using:

• evo for ATE/RPE
• Internal logging system for CPU, FPS, and memory

---

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.