- Mmnorm reconstructs complex hidden forms that use wi-fi frequencies without touching the object
- Robots can now see internal messy drawers using reflected signals of the surrounding antennas
- MIT technique exceeded the current radar accuracy in 18% in more than 60 tested objects
In environments where visibility is obstructed, such as internal boxes, behind the walls or under other objects, artificial intelligence could soon have a new way to get ahead.
MIT researchers have developed a technique called MMNOR, which uses millimeter wave signals, the same frequency range as Wi-Fi, to rebuild hidden 3D objects with surprising precision.
“We have been interested in this problem for quite some time, but we have been hitting a wall because the past methods, although they were mathematically elegant, did not take us where we needed to go,” said Fadel Adib, lead author of the study and director of the signal kinetics group at MIT.
Overcome radar limitations
The above techniques depend on the posterior projection, which produces low resolution and failure images when applied to small objects stamped as tools or utensils.
The researchers found that the failure lies in the supervision of a physical property known as specularity, how the reflections of the millimeter waves behave such as mirror images.
Instead of simply measuring where the signals are recovered, MMNorm estimates the surface address, what researchers call the normal surface.
“Trusting specularity, our idea is to try to estimate not only the location of a reflection in the environment, but also the direction of the surface at that point,” said Laura Dodds, the main author of the newspaper.
By combining many estimates of different antenna positions, the system reconstructs the 3D curvature of an object, distinguishing between forms as nuanced as the handle of a cup or the difference between a knife and a spoon in a box.
Each antenna collects reflections with variable force depending on the orientation of the hidden object.
“Some antennas could have a very strong vote, some could have a very weak vote, and we can combine all the votes to produce a normal surface that is agreed by all the antenna locations,” Dodds added.
This new approach achieved a 96% reconstruction precision in more than 60 objects, exceeding the existing methods that only reached 78%.
The system worked well in objects made of wood, plastic, glass and rubber, although it still fights with dense metal or thick barriers.
As researchers work to improve resolution and material sensitivity, potential use cases are growing.
In safety scan or military contexts, MMNorm could rebuild the shape of hidden items without opening bags or boxes.
This capacity could be essential for robots with AI in the automation of warehouses, search and rescue or even assisted life environments.
Through Techxplore
