A robot can infer a hidden object's shape, but that is not the same as seeing through a wall.
At a glance
MIT researchers used reflected millimeter-wave signals to reconstruct hidden objects in a research prototype. The method was tested on more than 60 objects.
Why this matters
Ordinary cameras need a clear line of sight. Radio signals can reflect around a scene, so a robot can collect clues even when an object is visually obscured. The MIT team's method uses those reflections to estimate a three-dimensional shape. That makes it interesting for cluttered warehouses, where a robot may need to find and grasp something a camera cannot directly observe.
MIT reports reconstructions of more than 60 objects and about 40 percent less error than the comparison approach in its tests. Those figures describe a research evaluation, not a guarantee for every material, room or industrial setting. The underlying work is about reconstructing shape from indirect signals and estimating where the object is; it does not produce a conventional photograph of the hidden object.
A practical question is whether the reconstruction is accurate enough for a robot to act on it safely. A rough estimate might help choose where to look or how to approach an object, while a delicate grasp may require more precise information. The next step for a real deployment would be to test reliability in the messy environments where the technique is supposed to help.
What to keep in mind
The improvement figure comes from the reported experiments. The method does not work through metal or very thick walls and does not establish that a production robot can identify every hidden object.
Read the source
Ahaia's short-form feed uses AI-assisted summaries. This expanded context was prepared separately from the cited source. Check the original before relying on a claim.