CA-SpaceNet: Counterfactual Analysis for 6D Pose Estimation in Space

Abstract

Reliable and stable 6D pose estimation of uncooperative space objects plays an essential role in on-orbit servicing and debris removal missions. Considering that the pose estimator is sensitive to background interference, this paper proposes a counterfactual analysis framework named CA-SpaceNet to complete robust 6D pose estimation of the space-borne targets under complicated background. Specifically, conventional methods are adopted to extract the features of the whole image in the factual case. In the counterfactual case, a non-existent image without the target but only the background is imagined. Side effect caused by background interference is reduced by counterfactual analysis, which leads to unbiased prediction in final results. In addition, we also carry out low-bit-width quantization for CA-SpaceNet and deploy part of the framework to a Processing-In-Memory (PIM) accelerator on FPGA. Qualitative and quantitative results demonstrate the effectiveness and efficiency of our proposed method. To our best knowledge, this paper applies causal inference and network quantization to the 6D pose estimation of space-borne targets for the first time. The code is available at https://github.com/Shunli-Wang/CA-SpaceNet.

Mobile-URSONet: an Embeddable Neural Network for Onboard Spacecraft Pose Estimation

“…Given that they use a Xilinx Ultra96 board, which contains a single-core Xilinx DPU on a small FPGA, achieving real-time inference may not be possible. The other FPGA Ultra96 implementation by [12] implemented only a single layer of their neural network on the FPGA, also hindering the calculation of energy efficiency. Our implementation is almost 9 times faster and over 38 times more energy efficient than the Intel Atom-based implementation by [6].…”

“…Previous research on real-time embedded SPE has made significant strides, as seen in works by [6], [13], [11], and [12], employing various hardware and quantization strategies. However, these studies often lack detailed methodology, particularly regarding quantization's effects on accuracy and efficiency, and frequently omit source code, hindering reproducibility.…”