The system of recognition of moving objects behind obstacles by means of Wi-Fi-echolocation appeared not yesterday, but now its service was assigned to a neural network. With the new algorithms, everything changed – there was a technology called RF-Pose. She does not “see” the person behind the obstacle, but can recognize who it is and what it does.
The principle of action has remained the same, the Wi-Fi signal when passing through obstacles changes its parameters, so if we know the bandwidth of the wall, then the rest of the data can calculate the location of objects behind it. RF-Pose works in two-dimensional space, but with exceptional accuracy, because AI builds an object model and analyzes its behavior, rather than simply measuring fluctuations in radio waves. It looks like a child's drawing of a man from wands, and to recognize the person, the personality of a person, the system is not able to. But where his gaze is directed – without difficulty.
Although AI with RF-Pose does not see a person behind the wall in a literal sense, algorithms for analyzing and predicting behavior allow him to draw a very reliable picture. Especially, if we give the neural network time and material for training, 100 people took part in the experiment, and by the end of the study AI confidently identified them “by walking” in 83% of cases. This approach helps to neutralize the lack of echolocation, minimize the effect of interference on the recognition of hidden people.
The authors of the development say that, given the proper skill level, the AI will be able to recognize what the offender is armed behind the wall, and whether he is ready to attack or simply hides. And also – to understand by gestures that a lonely patient in the ward urgently needs help, or collect dirt on negligent employees who sabotage the work process. Since the technical basis for applying RF-Pose is fairly simple, the technology has already been introduced into the category “not for free use”.js.src = “&version=v2.8”; 'script', 'facebook-jssdk'));