Improvement of organizational and technological solutions in robotic inspection and diagnostics of the technical condition of foundations and underground structures

Abstract

The article considers the improvement of organizational and technological solutions for robotic inspection and diagnostics of the technical condition of foundations and underground reinforced concrete structures in megacity conditions, where operational reliability is limited by the impact of aggressive groundwater, vibrations, and a shortage of safe access for personnel. It is shown that the transition from discrete manual control to automated monitoring is determined not only by the availability of mobile platforms equipped with ground-penetrating radars, ultrasonic, and optical-laser instruments, but also by the need for the legitimate integration of high-precision measurements into the practice of decision-making on strengthening and repair. Based on pilot implementations on 84 underground construction facilities from 2019–2024, a comparison of traditional non-destructive testing methods with data from robotic complexes was performed, scanning datasets totaling about 12 TB and over 500 test protocols were analyzed, which enabled a statistical characterization of errors depending on the kinematic scheme of the platforms (wheeled, tracked, walking) and environmental conditions. Key dependencies demonstrate a nonlinear decrease in the probability of detecting microdefects with increasing movement speed, especially for wheeled chassis on wet surfaces; walking systems provide the highest accuracy despite performance limitations. It has been established that the labor cost profile in robotization is inverted: works on ensuring access and field surveying are reduced, but office processing of point clouds and interpretation of thermograms become the dominant “bottleneck,” forming a demand for machine filtering algorithms and automatic defect classification. Economic analysis of integral costs accounting for depreciation, software, and error risks reveals the efficiency boundary: with easy access, the manual method is cheaper, whereas with difficult and particularly hazardous access, robotic diagnostics significantly outperforms. Additionally, the reliability of communication channels in underground spaces was quantitatively assessed, limitations of high frequencies in labyrinthine geometry and the compromise of low-frequency bands in terms of bandwidth were demonstrated, justifying hybrid communication schemes. Requirements have been formulated for updating the regulatory framework, metrological verification of integrated robotic systems, the legal status of the “digital replica,” machine-oriented defect classifiers, and unified protocols for data integration into facility information models, including the prospect of continuous monitoring and accumulation of signature databases for training neural network models and transitioning to predictive repairs.