Can robots independently repair defects on plastered walls? This research introduces Puttybot, a sensorized robot designed for autonomous putty plastering, particularly in transition areas where existing robots struggle. The Puttybot incorporates a mobile chassis, a lift platform, and a macro/micromanipulator. Its force-controlled scraper parameters dynamically adjust rigidity based on the putty applied, ensuring optimal performance. The system employs convolutional neural networks to identify plastering defects and relies on force feedback to execute precise repairs. Calibration of force-controlled scraper parameters allows dynamic modification of rigidity in response to the applied putty. During autonomous plastering trials, Puttybot effectively processed a large-scale wall without human intervention. The effectiveness of Puttybot was validated through autonomous plastering trials that involved processing a large-scale wall without human intervention. This innovative approach promises enhanced automation and consistent quality in plastering, addressing the limitations of current robotic solutions. This technology has implications for robotics, automation, and construction.
This article in the Journal of Field Robotics fits squarely within the journal's scope, focusing on advancements in robotics for real-world applications. The Puttybot project exemplifies the journal's emphasis on innovative robotic systems capable of autonomously performing complex tasks in unstructured environments. The research aligns with the journal's interest in robotics contributing to sectors such as construction and automation.