Overall System Depiction
Our system performs autonomous reforestation using tree seedlings. The user selects the planting area and seedling types, after which the robot autonomously plants the seedlings and reports planting progress back to the user.
On the technical side, localization is handled via GNSS RTK. The system determines target planting spots within the selected area, plans cost-efficient paths to reach them, and navigates while performing real-time obstacle avoidance. Upon arrival, the robot executes a three-step planting process — digging a hole sized to the seedling’s rootball, depositing a seedling from onboard storage, and tamping the soil — with seedling reloading occurring in parallel to maximize throughput. If digging fails, the location is marked unsuitable and the planner redirects the robot to the next spot.
After each successful planting, the coordinates are added to a dynamic site map that is both reported to the user for progress monitoring and fed back into path planning to prevent future paths from crossing planted seedlings.
