Robotic Floor Nailing of a Modular Container Building Structural Assembly Based on Multilevel Localization and Active Planning

The existing floor nailing process is labor intensive and inefficient and relies on manual operations, including marking lines, drilling holes, nailing self-tapping screws, and smoothing surfaces. No dedicated automated solution exists for floor nailing, and localization methods capable of meeting the precision requirements of nailing tasks in constrained environments remain underdeveloped. Additionally, existing planning approaches lack deep integration of real-time sensor feedback with actuator control, limiting their effectiveness in precise execution. To address these challenges, this study introduces a robotic solution for automated floor nailing. The need for repetitive mapping is circumvented by implementing a multilevel localization method for the nailing robot. Utilizing design information, the coarse localization can ascertain the robot’s location within an unfamiliar modular container building. Based on active perception, refined localization further enhances the accuracy of the nailing points. Given effective localization, the robot is capable of autonomous navigation through active planning, which is based on boundary detection and path tracking. Experimental results demonstrate the feasibility and robustness of the proposed robot under various nailing conditions, ensuring its applicability in modular container building structural assembly. Furthermore, the proposed multilevel localization algorithm and Backstepping-based active planning method effectively enhance navigation accuracy and achieve millimeter-level nailing precision.