High-speed wireless signals come with a flaw: they can’t penetrate walls. As engineers turn to sub-terahertz frequencies to support the massive data needs of virtual reality, AI, and autonomous vehicles, even minor obstructions—like a bookcase or a passing person—can disrupt connectivity.

A team at Princeton University has now developed a system that enables these ultra-fast signals to bend around indoor obstacles, ensuring uninterrupted data flow even in cluttered environments.

Figure 1. Wireless Airy Beams Bend.

The approach combines physics with machine learning to generate “Airy beams”—curved transmission paths that can steer around objects rather than simply reflecting off them. First theorized in 1979, Airy beams were mostly studied for their physics, but Princeton researchers extended the concept by training a neural network to select and adapt the optimal beam in real time as obstacles appear or move. Figure 1 shows wireless airy beams bend.

“As our world becomes more connected and data-hungry, the demand for wireless bandwidth is soaring,” said Yasaman Ghasempour, assistant professor of electrical and computer engineering at Princeton. “Sub-terahertz frequencies can unlock much greater speed and capacity.” She noted that the work could pave the way for wireless systems capable of carrying up to 10 times more data than today’s networks.

Signals That Adapt on the Fly

Unlike low-frequency radio waves, sub-terahertz beams travel in narrow paths, making them vulnerable indoors. Traditional solutions, such as reflectors, are impractical in real-world scenarios [1]. By shaping the signal itself to curve—like a curveball in baseball—Airy beams offer a way to maintain connectivity without additional infrastructure.

Graduate student Haoze Chen, the study’s lead author, explained: “Most work with Airy beams has focused on creating them and studying their physics. What we’re doing is generating and optimizing them—choosing the right beam for the situation. You want the link to adapt when there’s no line of sight.”

Virtual Training for Real-World Networks

To make this possible, the team built a simulator to train the neural network across countless indoor layouts, avoiding the need for impractical physical tests. “For Airy beams, there are infinite ways to curve, depending on how and where the bend happens,” Chen said. “A transmitter can’t scan through them all.”

Instead of brute-force training, the researchers integrated physics principles into the neural network, dramatically improving its efficiency. Once trained, the system quickly adapts to changes, sustaining strong connections in dynamic environments.

Experimental tests in complex indoor setups confirmed the technology’s potential. While still in the research stage, the results suggest that Airy beams could become a practical solution for next-generation wireless systems, enabling robust, high-capacity connectivity in the sub-terahertz spectrum.

Reference:

1. https://interestingengineering.com/science/princeton-wireless-airy-beams

Cite this article:

Keerthana S (2025), Wireless Airy Beams Bend Around Indoor Obstacles to Deliver Seamless High-Speed Connectivity, AnaTechMaz, pp.204