aselimc Sensor Fusion Pipeline
EKF State Estimation (robot_localization)
# ekf.yaml
frequency: 50.0
odom0: /wheel_odom
odom0_config: [true, true, false, false, false, true, ...]
imu0: /imu/data
imu0_config: [false, false, false, true, true, true, ...]
Calibration
- Camera-LiDAR: extrinsic calibration with checkerboard or targetless methods
- Camera-IMU: use kalibr for spatiotemporal calibration
- Hand-eye: AX=XB solvers (OpenCV, easy_handeye)
Point Cloud Registration
import open3d as o3d
result = o3d.pipelines.registration.registration_icp(
source, target, max_distance=0.05,
estimation_method=o3d.pipelines.registration.TransformationEstimationPointToPlane())
Key Libraries
robot_localization, Open3D, kalibr, kiss-icp