Unstaffed heavy machines are used to build a huge dam in Higashi-Naruse, Akita Prefecture. (Video by Yoko Masuda)
HIGASHI-NARUSE, Akita Prefecture—Two green bulldozers smoothly spread a mixture of pebble and cement at the construction site of a mountainous dam, followed by four vibratory rollers that compacted the fresh layer into place.
By design, no one was at the wheel of any of these heavy machines.
Such autonomous equipment had previously been a rarity on construction sites. But with worker shortages hitting the industry, general contractors are now welcoming the technology.
And they have quickly learned the advantages of deploying tireless, multi-tasking “workers” adept at performing Herculean tasks with speed and precision.
Autonomous building system A4CSEL (Quad Accel), developed by leading general contractor Kajima Corp., has been used at the Naruse Dam in Higashi-Naruse, a village with a population of 2,300, since 2020.
Up to 14 unmanned machines have been operating simultaneously since then, quietly building most parts of the massive embankment.
Civil engineering work entered the most important phase in summer 2024. And artificial intelligence (AI) plays a central role in the system.
For instance, AI determines the optimal operational pattern using simulation software based on human workers’ control of heavy machinery.
Combining as many as 7,000 micro-tasks, AI suggests the most efficient approach to using multiple vehicles simultaneously.
The self-driving heavy machines can operate around the clock, performing work that would otherwise require three shifts of human labor.
“It is impossible for humans to do all tasks evenly,” said Kenichi Hamamoto, a senior researcher at the Kajima Technical Research Institute. “Machinery can do it, though.”
What role should humans take under A4CSEL? A short distance from the dusty building site, three employees in business suits were watching the vehicles’ performance on a monitor inside a control room.
DIMINISHING LABOR POOL
The engineering and construction industry had been slow to develop techniques and technologies to make up for staff shortages in maintaining infrastructure facilities.
Although the automation of building procedures emerged during Japan’s late-1980s asset-inflated economic boom, its development did not gain much momentum due to deep-seated concerns about potential job losses in rural areas.
This all changed after the labor crunch became too serious to ignore amid the dwindling birthrate and graying of society.
“Before machinery could deprive us of our jobs, construction projects themselves appeared to become impossible to proceed with,” Hamamoto recalled.
The number of workers in the construction industry peaked at 6.85 million in 1997, according to infrastructure ministry statistics.
The figure has fallen by 30 percent to 4.79 million in 2022, and is expected to decline by an additional 20 percent by fiscal 2040 from fiscal 2020.
The industry’s image associated with heavy labor and dangerous tasks has discouraged young people from entering the field.
Currently, one in four skilled workers in the construction industry is 60 or older, and most of them are set to retire within the next decade.
Seeing AI begin to play an indispensable part in automation and robotics, Kajima started developing its own autonomous construction technology in 2009.
The contractor’s technology was first put into practical use in 2015 at a dam construction project in Fukuoka Prefecture.
For the Naruse Dam in Akita Prefecture, AI decided the most effective workflow.
To complete an embankment, many 75-centimeter-thick layers of cement and other materials must be formed across the 755-meter-long dam site. It takes 70 hours to press and finish one layer.
The monthly construction output of the 14 vehicles operating simultaneously at the Naruse Dam surpassed the previous national record for dam development projects.
