Russian scientists have proposed using neural networks to identify hidden defects in the roadway that require repair. In this respect informs TASS, referring to Maxim Bartenev, head of Poddorozhny Hardware and Software Complex LLC.
Not all road defects are visible to the naked eye. Water may wash a void in the ground or the foundation of the road may shake a lot. If such problems are noticed in advance, they can be fixed with the help of scheduled repairs. In particular, it is possible to “look” several meters below the roadbed using a ground-penetrating radar, a device that sends radio waves to the ground and analyzes the received response.
“In a single set, we will use neural networks to analyze large databases taken by deep georadar up to 200 km per day from roads. Tracking roads to a depth of 30-40 m with the help of our development will help to quickly identify areas of watering, voids, potholes, i.e. areas that are currently problematic or may be problematic in the future. Artificial intelligence should help us predict the evolution of these processes. “The prototype of a special trailer-shaped car should be ready in November, we plan to receive the first data with the results in March next year,” he said.
For scanning, a trailer with radar will need to travel along the road at a speed of 40 km / h, it can be towed using a car. The working height of the radar on the canvas is 2 centimeters. As a result, road workers will have large amounts of data to be analyzed using a neural network. An algorithm based on machine learning will predict how long it will take for the structure to fail. According to the director of the company, the pilot implementation is planned in the Moscow region in the near future.
Formerly paleontologists to solvethat the ancient megalodon shark was warm-blooded.