- Статьи
- Science and technology
- On the digital eye: a new construction robot turned out to be ten times more efficient than a human
On the digital eye: a new construction robot turned out to be ten times more efficient than a human
Russia has developed an autonomous robot with AI for monitoring construction sites. The device moves around the site on its own, takes a circular survey and generates a digital model, after which the system compares it with the project and work schedule. This allows you to identify defects and deviations from the plan, as well as estimate the amount of work performed without manually analyzing a large number of photos. The stated productivity of the robot reaches 15 thousand square meters in two hours. According to experts, when performing comparable tasks, such a system can examine 60 times a larger area per unit of time than a specialist during a crawl. However, the robot has yet to prove its stability in the constantly changing conditions of a real construction site.
A robot instead of a crawler
Russian scientists have created a robotic solution for digital construction management. The system is based on the ADCM platform, which collects data from the site using 360-degree photography, analyzes it using AI, and compares the actual state of the facility with an information BIM model - a three—dimensional digital twin — and a work schedule.
The robot complements the system with automated information collection directly on the construction site. He moves around the object on his own and regularly records its condition. In two hours, the device is capable of examining from 10,000 to 15,000 square meters, the authors of the project told Izvestia. Thus, data on the progress of construction can be obtained regularly, and not only during scheduled rounds. The system compares new results with previous measurements and design data, which makes it possible to identify deviations and changes in the facility faster.
The declared productivity of the robot reaches 7.5 thousand square meters per hour. To assess the effectiveness, this indicator can be compared with the results of a manual site survey. If a specialist manages to thoroughly examine, photograph and initially analyze about 1 thousand square meters in an eight-hour working day, his productivity is approximately 125 square meters. m per hour. In this case, according to experts, the robot is able to survey a comparable area 60 times faster than a human.
— At the same time, the 60-fold advantage is a calculated estimate, rather than the result of a direct comparative test of a robot and a human. The actual productivity of a specialist depends on the complexity of the object, the route of the bypass, the number of necessary images and the requirements for fixing defects. The more detailed the manual control should be, the more time it takes to inspect the same area," explained Nikita Kulikov, Director General of the NGO PravoRobotov, Candidate of Law.
For developers and general contractors, such automation is especially relevant at large facilities, where engineering services have to regularly monitor tens of thousands of square meters, the expert believes. In manual work, some of the data is collected during the rounds, after which they have to be separately systematized and compared with the project documentation. Automation allows you to combine these operations into a single digital circuit.
The project received a grant from the Innovation Promotion Foundation and was among the 25 most promising from Russia, India and South Korea in the seventh wave of the Sber500 accelerator. For the first time, the development was presented at the demo day of the international accelerator within the framework of the Moscow Startup Summit.
From control to robotics
According to experts, automation of construction control can solve one of the key problems of large facilities — the fragmentation of work progress data. Each group of specialists is usually responsible for its own area, which is why deviations at the junction of several zones may go unnoticed.
— Automation of construction control is one of the most practical areas for the application of robotics and artificial intelligence. For the customer, this means early detection of defects and schedule delays, and for the contractor, fewer disputes about the volume of work performed," said Evgeny Vishnevsky, NTI expert on new materials and technologies.
According to him, the next important stage for developers will be to go beyond the closed pilot project. The real construction site is constantly changing: new obstacles, equipment and temporary structures appear on it. Therefore, the system needs to maintain stable operation in such conditions. In addition, the effectiveness of the technology directly depends on the relevance of the digital model of the object, with which the robot compares the results of the survey.
Another potential effect is related to fixing defects and the volume of work performed. If data on the condition of the facility is collected regularly and compared with a digital model, it will be easier for the customer and contractor to determine when the deviation occurred and assess its scale.
— The issue of construction control is always very acute during the construction of any facility. Each group of workers, as a rule, is responsible only for its own block and area of responsibility, which can lead to missing the whole picture, and this can lead to a delay in the construction of the entire facility. Monitoring of the construction site will allow you to see the implementation of the project in real time," said Nikita Kulikov.
At the same time, the expert drew attention to the limitation of the technology: visual inspection is not enough to fully monitor all processes.
— A survey of the external contours alone will not provide a complete picture of the project, but it will undoubtedly facilitate and accelerate the work of foremen to monitor the progress of construction, as well as strengthen independent control. This will give the customer additional advantages. Such monitoring is especially important in the implementation of government contracts," Nikita Kulikov added.
The claimed development capabilities require additional verification, said technology expert Alexander Smirnov.
— It is important to separately evaluate the quality of navigation on a changing construction site, the accuracy of shooting and comparing data with the project, as well as the ability of the system to independently identify defects, and not just record changes, — he said.
The prototype of the inspection robot is currently being tested. In parallel, the company develops solutions for painting, marking and drilling. In a separate area, the company cooperates with Uzbekistan in the field of AI and robotics for construction. In the future, the data collected on the construction site can be used not only for monitoring, but also for setting tasks for robotic devices directly on the site.
Переведено сервисом «Яндекс Переводчик»
