Scholarlink

<p>As planetary and deep-space exploration missions grow in complexity, precise autonomous navigation during descent and landing becomes increasingly critical. Traditional inertial navigation systems (INS) accumulate errors over extended descent phases, often resulting in significant landing errors. Terrain relative navigation (TRN) has emerged as a robust alternative, allowing spacecraft to localize by comparing onboard sensor observations with preloaded reference maps. This article presents a comprehensive review of advances in vision-based TRN, with a focus on three core approaches: template matching, pattern matching, and deep learning–integrated approaches. We conduct a consistent comparison within and across these approaches, emphasizing their respective strengths, limitations, and mission-specific applicability. Based on these observations, we identify current challenges and offer forward-looking recommendations to support future research in vision-based TRN. A key trend is the increasing integration of deep learning to enhance robustness and accuracy under diverse conditions, despite higher computational demands. This review provides a structured understanding of current vision-based TRN developments and outlines future directions for enhancing reliability, precision, and adaptability in planetary landing missions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Review of Vision-Based Terrain Relative Navigation for Planetary Descent and Landing of Space Vehicles

  • 期刊论文

  • 发表日期：

    2026年2月23日

查看全文
  • Mohomad Aqeel Abdhul Rahuman,
  • Kyuman Lee

摘要

As planetary and deep-space exploration missions grow in complexity, precise autonomous navigation during descent and landing becomes increasingly critical. Traditional inertial navigation systems (INS) accumulate errors over extended descent phases, often resulting in significant landing errors. Terrain relative navigation (TRN) has emerged as a robust alternative, allowing spacecraft to localize by comparing onboard sensor observations with preloaded reference maps. This article presents a comprehensive review of advances in vision-based TRN, with a focus on three core approaches: template matching, pattern matching, and deep learning–integrated approaches. We conduct a consistent comparison within and across these approaches, emphasizing their respective strengths, limitations, and mission-specific applicability. Based on these observations, we identify current challenges and offer forward-looking recommendations to support future research in vision-based TRN. A key trend is the increasing integration of deep learning to enhance robustness and accuracy under diverse conditions, despite higher computational demands. This review provides a structured understanding of current vision-based TRN developments and outlines future directions for enhancing reliability, precision, and adaptability in planetary landing missions.