LLM-Integrated Autonomous Mobile Robot
ROS 2 navigation, SLAM, state estimation, and voice-controlled autonomy
I contributed to the development of a mobile robot prototype that combined autonomous navigation, perception, and large-language-model integration. The system was designed to support natural-language command execution in indoor environments.
Main work
- Designed an AMR chassis in CAD.
- Deployed ROS 2 Nav2 with LiDAR-based SLAM for real-time indoor mapping and navigation.
- Fused wheel odometry, IMU, and magnetometer data using a Kalman filter for robust state estimation.
- Integrated the Gemini LLM API with voice recognition and TTS for natural language–commanded control.
Impact
This project showed how autonomous mobile robots can combine classical navigation stacks with modern language interfaces to make human-robot interaction more intuitive.