Terrain Modeling and Cost Map Construction for Autonomous Agricultural Vehicles in Hilly Orchards: A Review

Area of Science:

• Agricultural Robotics
• Geomatics Engineering
• Environmental Modeling

Background:

• Autonomous navigation in hilly orchards faces challenges from rugged terrain and dense canopies.
• Existing environmental modeling techniques often neglect the propagation of elevation uncertainty during Digital Elevation Model (DEM) reconstruction.
• This oversight critically impacts terrain risk assessment and navigation strategies for agricultural vehicles.

Purpose of the Study:

• To review the propagation of elevation uncertainty from sensor data (RTK-GNSS, LiDAR, computer vision) through the DEM reconstruction pipeline.
• To analyze the impact of DEM errors and vertical inaccuracies on terrain analysis and vehicle navigation safety.
• To identify practical challenges in hilly orchards and propose future research directions.

Main Methods:

• Error-propagation analysis of uncertainties from RTK-GNSS, LiDAR, and computer vision.
• Examination of how these uncertainties affect DEM reconstruction, feature extraction, and cost map generation.
• Assessment of the influence of DEM elevation errors on slope, roughness, and traversability calculations.

Main Results:

• Uncertainties from sensor data significantly propagate through DEM reconstruction and subsequent navigation tasks.
• DEM elevation errors directly impact the accuracy of slope estimation, roughness representation, and traversability assessment.
• Vertical inaccuracies in DEMs compromise vehicle stability and overall navigation safety in complex terrains.

Conclusions:

• Addressing elevation uncertainty is crucial for reliable autonomous navigation in challenging agricultural environments.
• Future research should focus on multimodal sensor fusion, uncertainty-aware modeling, and adaptive mapping strategies.
• Learning-based approaches for traversability assessment show promise for enhancing navigation safety and efficiency.