Utilization of FPGA for Onboard Inference of Landmark Localization in CNN-Based Spacecraft Pose Estimation

Kiruki Cosmas,Asami Kenichi

doi:10.3390/aerospace7110159

Kiruki Cosmas, Asami Kenichi

Open Access

https://doi.org/10.3390/aerospace7110159

Copy DOI

Journal: Aerospace	Publication Date: Nov 5, 2020
Citations: 19	License type: CC BY 4.0

Affiliation: Kyushu Institute of Technology

Abstract

In the recent past, research on the utilization of deep learning algorithms for space applications has been widespread. One of the areas where such algorithms are gaining attention is in spacecraft pose estimation, which is a fundamental requirement in many spacecraft rendezvous and navigation operations. Nevertheless, the application of such algorithms in space operations faces unique challenges compared to application in terrestrial operations. In the latter, they are facilitated by powerful computers, servers, and shared resources, such as cloud services. However, these resources are limited in space environment and spacecrafts. Hence, to take advantage of these algorithms, an on-board inferencing that is power- and cost-effective is required. This paper investigates the use of a hybrid Field Programmable Gate Array (FPGA) and Systems-on-Chip (SoC) device for efficient onboard inferencing of the Convolutional Neural Network (CNN) part of such pose estimation methods. In this study, Xilinx’s Zynq UltraScale+ MPSoC device is used and proposed as an effective onboard-inferencing solution. The performance of the onboard and computer inferencing is compared, and the effectiveness of the hybrid FPGA-CPU architecture is verified. The FPGA-based inference has comparable accuracy to the PC-based inference with an average RMS error difference of less than 0.55. Two CNN models that are based on encoder-decoder architecture have been investigated in this study and three approaches demonstrated for landmarks localization.

Highlights

The determination of the orientation and position of a spacecraft is an important aspect in various space operations
In this paper we present how such Convolutional Neural Network (CNN)-based approaches can be realized on resource constrained spacecrafts such as small satellites and space robotics
Since the ultimate goal of this work was to investigate the acceleration of the CNN-part of pose estimation algorithms on Field Programmable Gate Array (FPGA), it is essential that the network layers and operations are supported by the Deep Learning Processing Unit (DPU)

Summary

Introduction

The determination of the orientation and position of a spacecraft is an important aspect in various space operations. The increased space activities have led to fears of an increase in space debris [2]. To address such concerns, studies for space debris removal have been conducted [3,4,5], in which effective pose estimation is required. Nanjangud et al in [8] present the use of robotics and AI for on-orbit operations (O3) while using small satellites. They highlight some technical challenges in robotic O3, such as: pose estimation of chaser and target; autonomous rendezvous and docking manoeuvres

Objectives

Methods

Results

Discussion

Conclusion