TH Köln investigates the use of AI in project transports

The transport of mobile cranes for construction projects or rotor blades for wind turbines poses challenges for companies and authorities. To ensure the accessibility of critical sections of routes, such as bottlenecks, simulations using software are often employed. TH Köln is investigating in a feasibility study how this verification can be carried out more easily and time-efficiently, among other things, with the help of artificial intelligence.

(Köln) “Even the planning of transports is a complex undertaking. To obtain an exception permit, applicants must consider road conditions, bridges, tunnels, tight curves, inclines, declines, and obstructive vegetation, among other factors. To identify all critical points along the route, large amounts of data must be collected, processed, and visualized. Various research projects are underway at the university. Our goal is to develop and test a storage- and computation-efficient method for accessibility analysis,” reports project leader Dr. René Degen from the CAD CAM Center Cologne at TH Köln.

Comparing two approaches

At the beginning of the project, two real bottlenecks on a transport route will be surveyed using two different methods. In the first method to be tested, researchers will fly a drone along the transport route. The captured image data will be processed into a 3D point cloud using photogrammetry to create a digital representation of the bottlenecks. “Each image is tagged with a geotag indicating the drone’s location. In photogrammetry, software analyzes the photos to identify common features or points. By comparing these points across the various images, the software can determine the geographical position in 3D space,” explains Degen.

As a comparative method, the researchers at TH Köln are developing a mobile mapping system that is mounted on the roof of a vehicle. Each measurement point captured by a laser scanner is assigned a geographical position via satellite. The geodata is then merged into a 3D point cloud. “Additionally, we mount a camera on the measurement system and overlay the images captured with our software onto the point cloud, which is then colored according to the surroundings. This creates 3D model maps that are intended to facilitate visual verification of the recordings for users,” explains the project leader.

Route planning and collision checking

In addition to environmental data, information about the geometry of the large equipment to be transported is fed into an algorithm to train an artificial intelligence. “The algorithm is tasked with finding a route with the lowest possible collision risk. In the case of a bottleneck, for example, contact with shrubs or bushes is classified as less severe, while contact with a bridge or building wall is considered a potentially serious event. We then check which data basis of the two methods allows for a more reliable and simpler analysis,” says Degen.

To ensure practical relevance in the project, a research-accompanying working group consisting of companies, authorities, and associations is being established. Through continuous exchange with partners, it is to be ensured that the prototype data processing and simulation of accessibility meets acceptance among potential users such as transport companies, planners, and permitting authorities.

The feasibility study is intended to lay the groundwork for further research in this area. Additionally, it is planned to validate the results in a follow-up project and to transition them into the development of a marketable product.

About the project

The project “pointCloud2Model” is being conducted at the CAD CAM Center Cologne of TH Köln, led by Prof. Dr. rer. nat. Dr. h.c. Margot Ruschitzka. Dr. René Degen is taking over project management. The research-accompanying working group includes the Federal Association of Heavy Transport and Crane Operations e.V., the State Agency for Mobility Rhineland-Palatinate, the sensor provider IGI mbH, and the transport company BigMove AG. The one-year project is funded with 200,000 euros through the funding guideline “mFUND” of the Federal Ministry for Digital and Transport.

Photo: © CAD CAM Center Cologne