Purpose This study aims to present a façade-cleaning robot for safe, efficient maintenance of high-rise glass curtain walls, combining active safety with a block-wise cleaning strategy. Design/methodology/approach The robot uses a modular Cartesian architecture with a synchronous-belt lifting module, worm-gear self-locking and multipoint vacuum anchoring. A distributed master–slave controller coordinates multiaxis motion, and a block-wise planner decomposes façades into overlapping units. Indoor and outdoor experiments quantify safety, cleaning quality and efficiency. Findings Field tests with a 30 kg payload at 10 m show that the dual-channel active safety mechanism arrests motion in power failures and keeps the platform stable against disturbances. The block-wise strategy yields uniform coverage on façades with protruding frames and restores glass transmittance comparable to manual cleaning. The robot matches skilled workers in efficiency while offering higher consistency and about 26% lower water use. Research limitations/implications Validation covers only a single prototype on flat façades under moderate wind; long-term reliability, extreme weather and highly irregular façades require further study. Practical implications The robot provides a solution for safe, water-efficient maintenance of high-rise curtain walls, offering quantifiable safety and resource-usage indicators for facility managers. Originality/value Unlike existing façade-cleaning systems based on passive safety and linear scanning, this work integrates intrinsically safe lifting hardware, a dual active safety mechanism and a block-wise cleaning strategy into one deployable robot, providing a practical reference for intrinsically safe, resource-efficient façade maintenance.
Building similarity graph...
Analyzing shared references across papers
Loading...
S. K. Li
Chaofeng Liu
Zhiyuan Chen
Industrial Robot the international journal of robotics research and application
Shanghai University of Engineering Science
Building similarity graph...
Analyzing shared references across papers
Loading...
Li et al. (Wed,) studied this question.
www.synapsesocial.com/papers/69be36666e48c4981c67549a — DOI: https://doi.org/10.1108/ir-12-2025-0444