Researchers at the MIT Media Lab have demonstrated that consumer-grade LiDAR sensors can detect, track, and reconstruct hidden 3D objects around corners [1, 2, 3].
This development could significantly lower the cost of spatial perception for autonomous systems. By utilizing hardware already found in many high-end smartphones, the technology removes the need for expensive, specialized equipment to map unseen environments [2, 3].
Led by Siddharth Somasundaram, the team in Cambridge, Massachusetts, focused on the capabilities of smartphone-grade LiDAR [1, 2]. The system is designed to perceive objects that are not in the direct line of sight of the sensor — a task typically reserved for high-end industrial scanners [2].
The researchers achieved these results using a LiDAR system with a hardware cost of under $50 [3]. This price point makes the technology accessible for a wide range of applications beyond luxury electronics [3].
The team intends for this low-cost perception to benefit robots and autonomous vehicles [2, 3]. These machines often struggle with "blind spots" at intersections or corners, where a collision risk is highest [2].
Beyond industrial use, the researchers said there is potential for new consumer applications [2, 3]. The ability to see around corners could eventually be integrated into mobile apps or accessibility tools for the visually impaired [3].
“Consumer-grade LiDAR sensors can detect, track and reconstruct hidden 3D objects around corners.”
The transition of non-line-of-sight imaging from expensive laboratory equipment to sub-$50 consumer hardware represents a shift in how autonomous agents interact with their environment. If integrated into mass-market robotics or vehicles, this capability reduces the reliance on complex communication networks between devices to share blind-spot data, allowing individual units to independently perceive hidden hazards.
