Space debris pose and motion identification using 4D LiDAR

The advent of commercial off-the-shelf 4D LiDAR with ranging and velocimetry capabilities offers new opportunities in the field of spacecraft perception. This paper presents a method to estimate the position and orientation, as well as linear and angular velocities, of a space debris with unknown geometry using a 4D LiDAR. The method computes a bounding box from the range point-cloud, using the two techniques of principal component analysis and random sample consensus, to compute the debris size, orientation, and position of its centroid. The box orientation and its centroid position and velocity are then combined with the LiDAR speed measurements to compute the angular velocity of the space debris. The computations are passed into a consensus-based iterated sigmapoint Kalman filter to estimate the debris pose and motion states. The filter utilizes the computed orientation only when there is a consensus between the two box computation techniques. The method is verified through a simulated case study, running a specific scenario as well as Monte Carlo simulations, where a defunct satellite is identified by a commercially available 4D LiDAR onboard another satellite.