Adopted for Research and Development Subsidy by the Ministry of Land, Infrastructure, Transportation and Tourism. Toward Automating Bridge Painting with AI.

The number of bridges awaiting repainting is increasing, while the number of skilled painters is decreasing. Refining the painting process with digital twins and implementing real-world robots on site.

Painting bridges is still done by hand.

Standing on high scaffolding, entering dangerous areas, and skillfully applying paint. The skilled workers for this task are rapidly disappearing.

Hokuto Tekko Co., Ltd. (Headquarters: Kanazawa City, Ishikawa Prefecture, Representative Director and President: Yasuhide Koikeda, hereinafter "Hokuto Tekko") announces that the joint research "Research on the Development of an Automated Painting System for National Infrastructure Using Physical AI" has been adopted as a research topic under the "SBIR Construction Technology Research and Development Subsidy Program" (Fiscal Year 2028, Small and Startup Enterprise Type) by the Ministry of Land, Infrastructure, Transportation and Tourism. The co-researcher is Tengun-label Co., Ltd. (hereinafter "Tengun-label"), a startup based in Tokyo.

This research topic is one of the 17 newly adopted topics of this type, with a planned grant amount of 7,000,000 yen (7 million yen). The research representative is Koikeda Yasunori, a director of Hokuto Tekko.

▸ More bridges need repainting, but fewer people can paint them

Japan's stock of bridges is rapidly aging, with bridges over 50 years old since construction expected to increase significantly in the future. Increased corrosion damage and prolonged waiting times for repainting are making maintenance, centered on steel bridges, increasingly critical.

An even more serious issue is the lack of skilled workers. The construction industry as a whole is facing a shortage of skilled workers and an aging workforce, making it difficult to pass on the necessary skills for bridge painting. In the future, there is a high risk that painting operations themselves may become difficult to continue.

▸ Capturing the skills of craftsmen in the digital world

Recent developments in physical AI and digital twin technology make it possible to accurately reproduce the painting process in a virtual space and reflect the simulation results in the real world to achieve automation.

In this research, we aim to optimize the operation of painting robots and ensure quality in a digital twin environment, achieving low-cost and short-term development, and realizing on-site adaptive automated painting.

Expected effects

Significantly shorten the lead time from painting planning to construction

Shorten the work period on site, reduce rework, and stabilize painting quality

Ensure the safety of workers in high-altitude or hazardous painting operations

▸ Why is this possible?

A manufacturing company that knows the site and a startup with cutting-edge AI

Hokuto Tekko Co., Ltd. (Research Representative: Director Yasunori Koikeda)

Founded in 1934, Hokuto Tekko is a large-scale steel structure manufacturer headquartered in Kanazawa City, Ishikawa Prefecture. The company consistently handles the design, manufacturing, and construction of steel bridges and has accumulated essential base technologies for the long-term durability of bridge painting within the company. Recently, the company has also been involved in repair work for steel bridges in service.

Tengun-label Co., Ltd. (Co-researcher)

Headquartered in Shinjuku-ku, Tokyo, Tengun-label is a group of engineers specializing in AI, 3D image processing, sensing, and data analysis. The company is involved in the social implementation of digital twin technology and the analysis of 3D point cloud data and, as an NVIDIA Omniverse partner company, possesses the latest technology information in the field of physical AI.

A manufacturing company that knows bridges thoroughly and a group of engineers who capture the real world with point clouds and AI.

This combination is what allows us to aim for the realization of an automation system that actually works on site, not just theoretical.

▸ The business plan was born from regional collaboration between industry, academia, and government

The business plan for this research was not suddenly created.

The starting point was the dialogue with local companies and universities over a year through the Kanazawa City commissioned project "TENJO KANAZAWA."

The trigger was February 2025. Yasunori Koikeda, a director of Hokuto Tekko, participated in the launch of "Monozukuri OpenMic," a gathering of Ishikawa's manufacturing industry to discuss the next move. In April of the same year, at the second installment "DX Promotion and AI Utilization," Hokuto Tekko's own initiatives were presented as a case study.

The turning point was autumn 2025. Amidst TENJO KANAZAWA's successive introductions of universities and companies, the business theme of "automating bridge painting" took shape. During this process, Tengun-label, which possesses physical AI technology, and Kanazawa Institute of Technology, which is skilled in industry-academia collaboration, were introduced.

Tengun-label's headquarters is in Tokyo. However, the company has participated in the membership program of the Local Revitalization Research Institute at the Hakusan Campus of Kanazawa Institute of Technology in Hakusan City, Ishikawa Prefecture, and already has a track record of activities in Hakusan City. The cutting-edge technology we were looking for was not far away. The local venue brought these together.

Then, in January-February 2026, under the guidance of TENJO KANAZAWA, the business plan was developed and applied for SBIR, and this time it was adopted under the same system.

The adoption of the subsidy system for this research and development and the start itself are the results of industry-academia-government collaboration.

▸ Implementation Plan

In the F/S (survey and analysis) of fiscal year 2026 (Reiwa 8), we will construct a model that reproduces painting work in a digital twin environment. In the subsequent R&D (research and development) of fiscal years 2027-2028 (Reiwa 9-10), we will expand to robot-based automated painting simulation in stable indoor environments (factories) and the development of autonomous mobile robots based on the results.

*Note: The research topic adopted this time under the "SBIR Construction Technology Research and Development Subsidy Program" (Fiscal Year 2028, Small and Startup Enterprise Type) by the Ministry of Land, Infrastructure, Transportation and Tourism is for the F/S of fiscal year 2026.

▸ Comments

Hokuto Tekko Co., Ltd.

Representative Director and President Yasuhide Koikeda

We have been building bridges and painting them for many years. Although painting work is mundane, it is an important process that determines the lifespan of a bridge. However, the number of skilled workers with these skills is decreasing year by year, and we have long been concerned about who will protect the bridges of this country in 10 or 20 years.

This time, we are challenging the cutting-edge technology of physical AI, not just following the trend. We are incorporating the skills of craftsmen into the digital world to pass them on to the next generation. Even as a local manufacturing company,

Keyword: