GNSS/LiDAR-Based Navigation of an Aerial Robot in Sparse Forests

Area of Science:

• Robotics
• Autonomous Systems
• Environmental Perception

Background:

• Autonomous navigation in forests is difficult due to degraded Global Navigation Satellite System (GNSS) signals and numerous obstacles.
• Detailed pre-mapping of forest environments is often impractical for unmanned vehicles.

Purpose of the Study:

• To develop a complete autonomous navigation solution for aerial robots in sparse forests.
• To enable reliable localization and obstacle avoidance in environments with intermittent GNSS availability.

Main Methods:

• A state estimator fusing GNSS, Attitude and Heading Reference Systems (AHRS), and LiDAR-based odometry using tree trunks for feature matching.
• A robust adaptive fusion algorithm based on the unscented Kalman filter for state estimation.
• A motion control strategy integrating a vector field for directional guidance and an optimal probabilistic planner for obstacle avoidance.

Main Results:

• Demonstrated effective localization by merging diverse sensor inputs (GNSS, AHRS, LiDAR).
• Successfully implemented LiDAR-based odometry using tree trunks for relative vehicle movement estimation.
• Validated the integrated motion control strategy for directional movement and obstacle avoidance in a forest setting.

Conclusions:

• The proposed navigation system effectively addresses the challenges of autonomous aerial robot operation in forest environments.
• The fusion of GNSS, AHRS, and LiDAR odometry provides robust localization.
• The integrated motion control ensures safe and directed navigation, as shown in experimental results.