Physical AI
Autonomous Mobile Robot with AI Vision
ROS 2 autonomous mobile robot with SLAM, navigation, obstacle avoidance and AI vision.
Overview
Research and development of a ROS 2-based Autonomous Mobile Robot that senses its environment, estimates its state, creates and uses maps, autonomously navigates between predefined stations, responds to static and dynamic obstacles, uses AI vision to influence robot behaviour, executes waypoint-based missions, and presents real-time map, camera, sensor, health and mission information through an AMR control dashboard.
Research Focus
- ROS 2 autonomous mobile robotics
- Odometry and sensor fusion
- 2D SLAM and map-based localization
- Autonomous goal and waypoint navigation
- Global and local path planning
- Static and dynamic obstacle avoidance
- AI vision for person, object and marker perception
- Vision-based robot behaviour
- Simulation-to-physical robot deployment
- Robot safety, health monitoring and mission control
Capabilities
- Differential-drive mobile robot control
- Wheel encoder, IMU, LiDAR, distance-sensor and camera integration
- ROS 2 nodes, topics, services, actions, parameters and launch systems
- URDF robot description and TF/TF2 coordinate frames
- Wheel odometry and fused state estimation
- 2D SLAM map creation, saving and reuse
- Map-based robot localization
- Autonomous navigation to goals and named waypoints
- Global planning, local control and costmap-based navigation
- Static and dynamic obstacle detection and response
- Live camera streaming with AI detection overlays
- Person-aware and object-aware robot behaviour
- AprilTag or QR-based station confirmation
- Multi-waypoint mission execution with Home, Pickup, Inspection and Delivery stations
- AMR Control Center dashboard with map, pose, path, camera, battery, health and mission status
- Simulation, visualization, logging and ROS data recording/replay
- Physical emergency stop, software stop, speed limits and critical-sensor safe-state behaviour