See through the wall with artificial intelligence

The technology is combined with radar and can deliver an enormous amount of information in almost real time.

A new artificial intelligence experiment could save lives – and not just those of soldiers. BWI, which is the IT provider and systems house of the Bundeswehr, has been investigating and testing the possibilities in the field of algorithms capable of learning since 2018, and is curious about the concrete benefits of the use of artificial intelligence in military conflicts. The innovation experiment called “Seeing through walls with artificial intelligence” is in an advanced stage. “We successfully tested the technology and proved its general feasibility” – He told Heiko Reiter, one of the two project managers.

Reiter is responsible for the organizational and commercial part of the innovation experiment, Jan Riedel is the technical leader and mastermind of the experiment. The radar technology used is not new. It is used in construction, for example, to detect electrical or water pipes in walls. “This is a commercial application, for military approaches we could not test any other professional solution on the European market,” explained Riedel. BWI uses radar in combination with artificial intelligence to detect people behind walls made of different materials in their experiment.

Radar emits high-frequency waves in the form of short pulses. Obstacles reflect these radar pulses and send them back to the receiver. The position of the localized objects can be determined from the transit time and direction of the radar signals. If constant measurements are taken, the movement of objects can also be determined. The radar beams are harmless and can penetrate the various materials the walls are made of. One exception is reinforced concrete: radar waves cannot penetrate steel.


In FPS games, you only need to download a cheat to see through walls, in reality it is more difficult

In the BWI project, radar technology is combined with artificial intelligence. The intelligent algorithm can distinguish between three things: whether they are objects or people; if it is about people, how many are there; and what their current state of movement is, i.e. whether they are walking, standing or sitting. “For the experiment, the project staff demonstrated actions in front of the radar sensor for several days and thus recorded different situations with another sensor measuring system in order to store the sequences of people’s movements in the system and train the application,” emphasized Riedel. In a possible further step, the training would be extended to animals and children.

Subjects were recorded sitting, standing, lying down and walking. This resulted in hundreds of gigabytes of data. The innovation experiment is basically based on sensor data analysis and machine learning approach. If people are behind a house wall, for example, they reflect the radio data transmitted by this obstacle. The radar records the reflected data and merges them with the reference data, the recorded movements. Artificial intelligence is fed with reference and radar data. In this way, the machine learns how people move from millions of data. Combined with radar technology, people can be recognized behind obstacles. The application runs on a regular laptop, so it does not require a lot of computing power. But the latter is already necessary from the point of view of system development.

The technology can recognize whether it is arm or leg movements, chest or abdominal breathing, or even pulse. It recognizes where people are and where they are based on even the smallest movements.

The application of artificial intelligence can be imagined according to the following scenario. Bundeswehr special forces are deployed to free the hostages. This is a life-threatening situation for soldiers, hostages and criminals. In order to minimize collateral damage, the situation is ideally clarified as precisely as possible before the attack. In a ten-room building, for example, you cannot kick in all the doors at the same time. If the door with the people behind it is broken only seconds later, there is a risk that the hostages will be shot. Therefore, the radar system is first used to see which room people are in.

The app can detect up to three people. The coordinates can be used to determine where they are in the room. Since the application has the possibility to reduce the delay of data processing, the image is created in quasi-real time. Movement status shows whether someone is walking, standing, sitting or lying down. From this, tactical considerations regarding access can be derived. All forecasts help to better assess the situation.

The experiment was presented at the Federal Ministry of Defense, where it received positive feedback. “During the development phase, we collected basic data in laboratory conditions, with the help of which we can train the artificial intelligence system,” emphasized Riedel. In order to find out how useful the technology is in real-world conditions, a practical test was also carried out with a prototype in a real building. One of the potential users of the technology is the German armed forces. According to Reiter, the police will be able to use the development for hostage rescues, the German Federal Technical Aid Organization for rescuing buried persons or the mountain rescue service for searching for avalanche victims.


Source: SG.hu Hírmagazin by sg.hu.

*The article has been translated based on the content of SG.hu Hírmagazin by sg.hu. If there is any problem regarding the content, copyright, please leave a report below the article. We will try to process as quickly as possible to protect the rights of the author. Thank you very much!

*We just want readers to access information more quickly and easily with other multilingual content, instead of information only available in a certain language.

*We always respect the copyright of the content of the author and always include the original link of the source article.If the author disagrees, just leave the report below the article, the article will be edited or deleted at the request of the author. Thanks very much! Best regards!