Adaptive Neural Network-based Unscented Kalman Filter for Robust Pose Tracking of Noncooperative Spacecraft

TL;DR

This paper introduces an adaptive neural network-based Unscented Kalman Filter for robust pose estimation of noncooperative spacecraft, combining CNN-based image analysis with online noise tuning for accurate rendezvous navigation.

Contribution

It develops a novel adaptive UKF framework that integrates CNN pose estimation with online process noise adjustment, validated on a new comprehensive rendezvous dataset.

Findings

Achieves sub-decimeter position accuracy

Attains degree-level orientation accuracy

Demonstrates robustness on real robotic testbed images

Abstract

This paper presents a neural network-based Unscented Kalman Filter (UKF) to estimate and track the pose (i.e., position and orientation) of a known, noncooperative, tumbling target spacecraft in a close-proximity rendezvous scenario. The UKF estimates the target's orbit and attitude relative to the servicer based on the pose information provided by a multi-task Convolutional Neural Network (CNN) from incoming monocular images of the target. In order to enable reliable tracking, the process noise covariance matrix of the UKF is tuned online using adaptive state noise compensation which leverages a newly developed closed-form process noise model for relative attitude dynamics. This paper also introduces the Satellite Hardware-In-the-loop Rendezvous Trajectories (SHIRT) dataset to enable comprehensive analyses of the performance and robustness of the proposed pipeline. SHIRT comprises the labeled images of two representative rendezvous trajectories in low Earth orbit created using both a graphics renderer and a robotic testbed. Specifically, the CNN is solely trained on synthetic data, whereas functionality and performance of the complete navigation pipeline are evaluated on real images from the robotic testbed. The proposed UKF is evaluated on SHIRT and is shown to have sub-decimeter-level position and degree-level orientation errors at steady-state.