His research team MITdrawing inspiration from the famous classic video game Tetrissucceeded in building a simple radiation detector, which detects radioactive sources in a safe and efficient way.
Her team Mingda Li (responsible for the project) reports that the device uses machine learning algorithms to process the data and map the radiating sources using only four detector pixels (pixels or pixels: a unit of measurement for graphics. If I imagine an image on the computer as a mosaic, each tile is a pixel).
In areas where accelerators or other devices with radioactive materials are used, there is always the risk of radiation leakage into the environment. Those responsible for personnel safety should be able to map the sources used in the workplace in time and locate the leak.
At first glance, Li explains that many times to confirm the origin of radiation from a source, we need to get close enough to it. However, as we get closer to the source, the risk of exposure increases. The goal of people recording radiation is to have as little radiation exposure as possible.
Expensive and complex to manufacture
Because radiation propagating in space can interact with the objects and devices present there, it often becomes quite complex and expensive (in terms of construction) to detect the radiation from a safe distance.
In a typical detector, irradiance spatial distribution maps are constructed by observing the intensity patterns recorded in a 10 × 10 (10cm × 10cm) array of detector pixels.
The biggest difficulty is that the radiation can approach the detector from different directions, making it difficult to extract information about the origin of the radiation and the reason for its emission.
To determine where the radiation is coming from, the researchers use certain “filters” that isolate some of the detector’s pixels at a time so they can pinpoint the source and direction of propagation.
Artificial intelligence as a solution
To solve the above problems, Li’s team used machine learning algorithms which make it possible to extract more and better quality data.
Additionally by creating a series of detectors with arrays of four pixels, in shapes such as tetromino (the “squares” of Tetris), discovered that they could achieve the same accuracy as detectors with larger arrays of pixels (in number).
The researchers put their algorithms and Tetris detectors to the test at Lawrence Berkeley Lab. The detectors, orbiting around examined radioactive sources, constructed radiation source maps that predicted with great accuracy the location of the source as well as the direction in which it transported the radiation. Their research was even published in the journal Nature.
The researchers, through the successful measurements they extracted, are optimistic that in the future their device will be able to be used on a daily basis, in laboratories, for radiation measurements. Thus making the everyday life of researchers safer.
#Radiation #detector #inspired #Tetris
