ALIW-farm

This dataset comprises multi-sensor data collected from a custom unmanned ground vehicle (UGV) platform navigating through diverse unstructured agricultural environments, including open farmlands, papaya groves, and GPS-denied betel nut forests. Sensor Configuration: LiDAR: Livox Mid-360 with integrated IMU (Inertial Measurement Unit) Wheel Odometry: High-precision wheel encoders for linear and angular velocity measurements Extrinsic Calibration: Pre-calibrated spatial relationships between LiDAR-IMU and wheel odometry frames Environmental Coverage: The dataset captures challenging scenarios typical of agricultural robotics applications: Open farmland: Feature-sparse arable land with sparse geometric constraints Orchard environments: Papaya groves with dense canopy structures GPS-denied areas: Betel nut forests with severe signal occlusion Data Characteristics: High-frequency IMU measurements for state propagation Wheel odometry data providing velocity and scale constraints 3D LiDAR point clouds exhibiting non-uniform spatial distribution Severe motion disturbances and terrain vibrations representative of real-world farm operations Intended Use: This dataset is specifically designed for evaluating tightly-coupled multi-sensor SLAM algorithms (LiDAR-IMU-Wheel fusion) in geometrically degraded agricultural environments.It serves as a complementary benchmark to existing public datasets (e.g., CitrusFarm) by providing: Diverse environmental diversity across different crop types Explicit LiDAR-IMU extrinsic parameters Real-world validation scenarios for robustness analysis File Format: ROS bag format with synchronized sensor topics including IMU raw data, wheel encoder readings, and LiDAR point cloud streams.