By developing infrastructure-facility materials that are less prone to deterioration than conventional ones, we aim to reduce the workload required for facility maintenance and operations and to generate revenue providing the materials to other companies. Leveraging more than 50 years of operational experience with resin-concrete manholes, NTT has accumulated deep insight into deterioration mechanisms. By improving resin concrete, we will develop infrastructure materials that do not deteriorate over a semi-permanent timescale.
Point cloud data with three-dimensional coordinate information is used in fields such as surveying, mapping, the metaverse, and robotics, and new applications are also being explored. To further utilize point cloud data, we will exhibit technologies being studied at our laboratory-including a method for acquiring high-precision point cloud data at low cost, analysis techniques such as automatic detection, and a database system that processes the entire workflow from data acquisition to analysis.
This exhibition will showcase technology designed to help instructions be communicated and comprehended during on-site work, thereby enabling tasks to be smoothly executed. A key feature of this technology is its ability to infer the instructor's intended actions and convey instructions clearly using multimodal representations, such as hand movements, hand shapes, and textual information. The aim is to more accurately deliver instructions and enable on-site work that does not depend on the worker's skill level.
This exhibit will introduce our Unsafe Behavior Detection Technology, designed to improve safety during on-site operations.
The technology leverages generative AI to automatically identify workers' unsafe behavior by analyzing work-related videos/images together with textual safety regulations.
By systematizing features related to worker actions and object states-including those not explicitly described in the safety rules-the system aims to enhance recognition accuracy.
This technology enables rule violations in daily operations to be detected early and helps prevent serious accidents.
In the maintenance of bridge-mounted conduits, the declining labor force is making it increasingly difficult to sustain conventional visual inspections that rely heavily on manual work. Visual inspections require on-site assessments, and the evaluation results may vary depending on the inspector.
To address this, establishing an inspection technology based on vibration characteristics-quantitative physical parameters-can eliminate discrepancies in judgment and reduce the risk of overlooking abnormalities. In addition, by measuring vibrations remotely using optical fiber sensing, on-site workload can be further reduced.
In the future, we aim to expand this technology's scope of application and enable it to be more broadly used in bridge structures.
This exhibition will introduce our current initiatives and the mock bridge model used for verification experiments.
As metal-based services are being discontinued, the costs of maintaining and removing overhead metallic cables urgently need to be reduced. We aim to reduce the workload required for removing cables by safely cutting and detaching them from hardware without loosening the tension or needing a worker to climb a pole.
The aim of this study is to fundamentally reduce infrastructure maintenance costs by extending utility pole spans and thereby reducing the number of poles in future access networks without copper cables.
Specifically, this study presents experimental and simulation-based evaluations of vibration behavior in long-span cables, as well as a cable-raising method using large sag.
These approaches simultaneously resolve load imbalance and insufficient cable height from the ground associated with pole reduction.
When utility poles are installed using a pole setting vehicle, LiDAR acquired point cloud data are utilized to generate geometric models of movable poles as well as surrounding obstacles, including existing poles, cables, and other infrastructure. When the movable pole model representing the construction target is detected to be in close proximity to obstacle point clouds or their corresponding models, the system automatically issues a hazard alert to the operator. This approach enhances on site safety and contributes to mitigating workforce shortages in construction.
When installing utility poles with a pole setting vehicle, point cloud technology is employed to identify surrounding obstacles, and an optimal path is automatically generated that ensures the target pole is safely and efficiently manipulated. Furthermore, the system provides operational instructions that enable the heavy equipment operator to accurately follow the automatically generated trajectory, thereby allowing even inexperienced operators to install utility poles.