The journal Nature Computational Science published a scientific study on the ability of artificial intelligence technologies to generate spatial layouts for cities more efficiently than humans. The study indicates the ability of artificial intelligence - by relying on machine learning technology - to create designs for cities that can be navigated in a few fractions of time, estimated at about 15 minutes. Machine learning is a branch of artificial intelligence and focuses on creating systems that learn - or improve performance - based on data received by users or provided to them. This study comes within new applications in which artificial intelligence techniques are used to find solutions to civil and urban planning problems, including finding solutions to improve cities that have become crowded with vehicles and facilities. The researchers in this study developed an artificial intelligence system to address urban planning, access to services and green spaces, traffic levels, and ideal road planning and land use, in line with the concept of the city. The dream of living in a green city According to a report published by the Science Alert website, artificial intelligence can achieve the dream of living in a cool green city full of parks, pedestrian walkways, bicycle lanes, and buses that transport people to stores, schools, and service centers within minutes. The report adds that the dream of urban planning, which is summed up in the idea of the “15-minute” city, has become closer to reality, as all basic needs and services are within a 15-minute walking distance, which leads to improving public health and reducing vehicle emissions, and artificial intelligence can help. Urban planners can achieve this vision faster.
To achieve this vision, automation scientist Yu Cheng and his colleagues, at the Beijing National Research Center for Information Science and Technology in China, worked on new solutions to improve cities, which have become more crowded and more polluted.
The research team developed an intelligent system that can handle tedious computational tasks for urban planning engineers, and this system provided 50% higher efficiency with respect to the three main metrics, access to services, access to green spaces, and traffic congestion.
The beginning of the project
Cheng and his colleagues started with a small project, modeling their model on urban areas of just a few square kilometers (about 3 by 3 blocks, an area known as a “block”).
After two days of training, and using several neural networks, the system was able to visualize ideal road and land use plans, consistent with the “15-minute city” concept.
It is true that the first calculations took only two days of work, but the system has proven its effectiveness, which means its ability to work in larger areas, taking into account that designing an entire city will be more complex and take a longer time.
Scientists believe that using the application may not be comprehensive or comprehensive, as it is sufficient to limit its work to some steps during planning cities or residential neighborhoods.
The system proved its ability to work with greater accuracy and speed. What engineers needed to calculate in 50 or 100 hours, artificial intelligence was able to calculate in a few seconds.
Researchers involved in the study believe that sharing AI with work on the most time-consuming tasks in urban planning would free up planners and engineers to work on other, more challenging, human-centered tasks, such as public engagement and aesthetics.
Comments
Post a Comment