In context
Flexible surgical robots are increasingly used to navigate tortuous and narrow luminal pathways, but contact with surrounding tissue can obstruct progress and risk damage. While much research has focused on visual perception, embodied perception—using the robot's own shape to understand and respond to its environment—remained underexplored. This study, published in the Chinese journal Robot in 2025, addresses that gap by integrating shape sensing, interpretation, and decision-making for safer autonomous intervention.
What was reported
Researchers from the Shenyang Institute of Automation and Soochow University proposed an embodied morphological perception method based on fiber Bragg grating (FBG) sensors. The method reconstructs the robot's shape using a discrete arc assumption with multi-core FBG fibers. To handle signal noise, they developed an improved moving average filter (IMAF) that progressively extracts influence from neighboring arc states and smooths endpoints, outperforming conventional filters in simulations and physical tests.
For environmental understanding, a spatiotemporal blockage detection method identifies where the robot is constrained by comparing shape deviations over time. Based on this, an intermittent intervention strategy mimics human catheter manipulation: when a blockage is detected, the robot retracts slightly before advancing again. This approach, guided by an instantaneous global inconsistency index (IGII), aims to maintain compliance and reduce tissue stress.
Experiments validated the feasibility of the blockage detection and intervention methods in various scenarios, demonstrating the IMAF's superiority over existing filters and the overall effectiveness of the embodied perception framework.
Why it mattered
This work advances the autonomy and safety of flexible surgical robots by enabling them to perceive and interpret their own morphological state during intervention. The proposed methods could reduce reliance on external imaging, lower tissue damage risk, and improve the intelligence of robotic systems in minimally invasive procedures, with potential extensions to other confined-environment applications.
"赋予柔性手术机器人具身感知的能力，意味着其能实现从感知到认知，可以像人类一样（甚至超越人类）感知、理解交互状态并响应环境信息作出决策。"
Source: 《机器人》期刊 (robot.sia.cn) · Published 2025-08-19 · “腔道介入交互约束下的柔性机器人具身形态感知方法”