Scholarlink

<p>Coverage Path Planning (CPP) is an essential task in many robotic applications such as cleaning, exploration, and agriculture. Reconfigurable robots are capable of altering their structure thereby accessing narrow spaces which is unfeasible for fixed-shape robots. However, the existing CPP strategies for reconfigurable robots have only focussed on executing predefined motion patterns to transition from one morphology to another which is adverse to harnessing the maximum potential from the reconfigurable robots. This paper proposes a novel online CPP approach employing adaptive morphology transition for a reconfigurable robot which aids in accessing confined spaces and perform reconfigurations. The proposed system utilizes a global coverage path planner that performs boustrophedon motion and a local path planner which is a reinforcement learning-based framework for enabling morphology transition to move through narrow areas. Experimental results reveal that the proposed system is superior to existing CPP research on reconfigurable robots in performing coverage and reconfigurations in confined areas.</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

Enhancing coverage efficiency of reconfigurable robots in confined environments through adaptive morphology transition using reinforcement learning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月5日

查看全文
  • W. K. R. Sachinthana,
  • J. J. J. Pey,
  • S. M. Bhagya P. Samarakoon,
  • M. A. Viraj J. Muthugala,
  • Mohan Rajesh Elara

摘要

Coverage Path Planning (CPP) is an essential task in many robotic applications such as cleaning, exploration, and agriculture. Reconfigurable robots are capable of altering their structure thereby accessing narrow spaces which is unfeasible for fixed-shape robots. However, the existing CPP strategies for reconfigurable robots have only focussed on executing predefined motion patterns to transition from one morphology to another which is adverse to harnessing the maximum potential from the reconfigurable robots. This paper proposes a novel online CPP approach employing adaptive morphology transition for a reconfigurable robot which aids in accessing confined spaces and perform reconfigurations. The proposed system utilizes a global coverage path planner that performs boustrophedon motion and a local path planner which is a reinforcement learning-based framework for enabling morphology transition to move through narrow areas. Experimental results reveal that the proposed system is superior to existing CPP research on reconfigurable robots in performing coverage and reconfigurations in confined areas.