Monday, February 16, 2026

Robots that can see around corners using radio signals and AI

Penn researchers developed HoloRadar, a system that reconstructs hidden 3D spaces beyond robots’ line of sight.

University of Pennsylvania School of Engineering and Applied Science

HoloRadar in Action — image:
HoloRadar uses radio waves to see around corners, allowing it to detect people at T-shaped intersections like the one pictured here.
view more
Credit: Sylvia Zhang, Penn Engineering

Penn Engineers have developed a system that lets robots see around corners using radio waves processed by AI, a capability that could improve the safety and performance of driverless cars as well as robots operating in cluttered indoor settings like warehouses and factories.

The system, called HoloRadar, enables robots to reconstruct three-dimensional scenes outside their direct line of sight, such as pedestrians rounding a corner. Unlike previous approaches to non-line-of-sight (NLOS) perception that rely on visible light, HoloRadar works reliably in darkness and under variable lighting conditions.

“Robots and autonomous vehicles need to see beyond what’s directly in front of them,” says Mingmin Zhao, Assistant Professor in Computer and Information Science (CIS) and senior author of a paper describing HoloRadar, presented at the 39th annual Conference on Neural Information Processing Systems (NeurIPS). “This capability is essential to help robots and autonomous vehicles make safer decisions in real time.”

Turning Walls Into Mirrors

At the heart of HoloRadar is a counterintuitive insight into radio waves. Compared to visible light, radio signals have much longer wavelengths, a property traditionally seen as a disadvantage for imaging because it limits resolution. Zhao’s team realized that, for peering around corners, those longer wavelengths are actually an advantage.

“Because radio waves are so much larger than the tiny surface variations in walls,” says Haowen Lai, a doctoral student in CIS and co-author of the new paper, “those surfaces effectively become mirrors that reflect radio signals in predictable ways.”

In practical terms, this means that flat surfaces like walls, floors and ceilings can bounce radio signals around corners, carrying information about hidden spaces back to a robot. HoloRadar captures these reflections and reconstructs what lies beyond direct view.

“It’s similar to how human drivers sometimes rely on mirrors stationed at blind intersections,” says Lai. “Because HoloRadar uses radio waves, the environment itself becomes full of mirrors, without actually having to change the environment.”

Designed for In-the-Wild Operations

In recent years, other researchers have demonstrated systems with similar capabilities, typically by using visible light. Those systems analyze shadows or indirect reflections, making them highly dependent on lighting conditions. Other attempts to use radio signals have relied on slow and bulky scanning equipment, limiting real-world applications.

“HoloRadar is designed to work in the kinds of environments robots actually operate in,” says Zhao. “This system is mobile, runs in real time and doesn’t depend on controlled lighting.”

HoloRadar augments the safety of autonomous robots by complementing existing sensors rather than replacing them. While autonomous vehicles already use LiDAR, a sensing system that uses lasers to detect objects in the vehicles’ direct line of sight, HoloRadar adds an additional layer of perception by revealing what those sensors cannot see, giving machines more time to react to potential hazards.

Processing Radio With AI

A single radio pulse can bounce multiple times before returning to the sensor, creating a tangled set of reflections that are difficult to untangle using traditional signal-processing methods alone.

To solve this problem, the team developed a custom AI system that combines machine learning with physics-based modeling. In the first stage, the system enhances the resolution of raw radio signals and identifies multiple “returns” corresponding to different reflection paths. In the second stage, the system uses a physics-guided model to trace those reflections backward, undoing the mirror-like effects of the environment and reconstructing the actual 3D scene.

“In some sense, the challenge is similar to walking into a room full of mirrors,” says Zitong Lan, a doctoral student in Electrical and Systems Engineering (ESE) and co-author of the paper. “You see many copies of the same object reflected in different places, and the hard part is figuring out where things really are. Our system learns how to reverse that process in a physics-grounded way.”

By explicitly modeling how radio waves bounce off surfaces, the AI can distinguish between direct and indirect reflections and determine the correct physical locations of a variety of objects, including people.

From the Lab to the Real World

The researchers tested HoloRadar on a mobile robot in real indoor environments, including hallways and building corners. In these settings, the system successfully reconstructed walls, corridors and hidden human subjects located outside the robot’s line of sight.

Future work will explore outdoor scenarios, such as intersections and urban streets, where longer distances and more dynamic conditions introduce additional challenges.

“This is an important step toward giving robots a more complete understanding of their surroundings,” says Zhao. “Our long-term goal is to enable machines to operate safely and intelligently in the dynamic and complex environments humans navigate every day.”

This research was conducted in the Wireless, Audio, Vision and Electronics for Sensing (WAVES) Lab at the University of Pennsylvania School of Engineering and Applied Science, and was supported by the University of Pennsylvania.

Method of Research

Experimental study

Subject of Research

Not applicable

Article Title

Non-Line-of-Sight 3D Reconstruction with Radar

How HoloRadar Works [VIDEO] |

HoloRadar relies on compact and nimble scanning equipment, opening up real-world applications.

Credit

Sylvia Zhang, Penn Engineering

HoloRadar works by reconstructing 3D scenarios from the bounces of radio waves.

Credit

WAVES Lab, Penn Engineering

Open-source modular robot for understanding evolution

A cost-effective, customizable quadruped could help researchers discover the particular advantages related to the length and segmentation of animal limbs

University of Michigan

Photos of the robot

What is it about a cheetah's build that enables it to run so fast? What gives the wolf its exceptional endurance?

While these questions can be partly answered through animal experiments, many contributing factors can't be isolated from one another. Now, a new tool has arrived: a highly customizable, open-source robot design called The Robot of Theseus, or TROT, developed at the University of Michigan.

Named in homage to Greek philosophy's "Ship of Theseus," the robot is composed of commercially available motors and 3D-printed parts, which can be rearranged to take on a broad array of designs. The plans address several pain points for animal researchers who might be able to harness robotics for biomechanical experiments, as well as for roboticists seeking more task-specific designs. Assuming access to 3D printers, the cost in parts and materials is under $4,000.

"In paleontology, we can go back and look at bones, but it's really difficult to understand how these changes in limb proportion, or in range of motion, may have affected the way an animal can move. There have been some really great insights on this question from robots that each mimic one extinct animal very precisely," said Talia Moore, assistant professor of robotics with a background in evolutionary biology and corresponding author of the study in Bionspiration and Biomimetics. "But each robot took years to design and construct.

"I wanted to make a robot that could easily shapeshift into several different extinct species proportions, so that we could compare them, and see how the evolution of those limb lengths and other features would affect their locomotion. With TROT, 60 million years of evolutionary changes in body size can happen in 20 minutes."

Usable, customizable and easy to measure

The modular robot plans and assembly guides offer three major benefits. First, they are usable by people without robotics degrees, with help from equipment that is available at many universities. As Moore pointed out, robotics offers insights into biological questions, but not many evolutionary labs have the benefit of robotics expertise.

Second, the robot's shape is highly customizable. While the published study focuses on four-legged designs, experimenters can change nearly any body segment—adding and removing parts, changing the range of motion and more. This means TROT can model most mammals and enable direct comparisons of variations on the same structure—for instance, between closely related extant and extinct species. And they can try out theoretical designs to determine whether they are biomechanically unfavorable or just untried by evolution.

Third, researchers mimicked the springiness and stiffness of muscular structures without actual springs or elastics, which can muddy measurements. TROT simulates this biological energy storage and return mechanism with backdrivable motors, which recover energy as they are driven backwards.

"Traditional robots are designed with an emphasis on industrial applications and are expensive to make. TROT was designed with ease of fabrication in mind," said Karthik Urs, a recent master's graduate in robotics and first author of the study.

"The overall part count is kept low, and most of the parts only fit together one way. That means that scientists can make most of the robot parts in-house with commodity 3D printers, assemble them and get to experimenting faster. It also makes the iteration process quick—key to enabling exploration in both robot and experimental design."

Isolating biomechanical factors that are tough to measure in animals

Moore was first inspired to make this robot when reading a 1974 experiment on running cheetahs and goats. Because the leg swings from the hip like a pendulum, physics holds that legs with more mass away from the hip, known as a greater moment of inertia, require more energy to redirect than legs that weigh the same but have most of the mass near the hip. This concept has informed the interpretation of evolutionary changes in legs—increasingly tapered limbs are likely associated with more efficient running.

However, the 1974 experiment showed that although a cheetah has a more favorable moment of inertia in its limbs, running costs nearly the same amount of energy as it does for a goat. Because so much else was different between these animals, Moore explained, the benefit from a lower moment of inertia was basically unmeasurable. In contrast, Moore's group varied only the weight distribution in their robot's limbs and was able to isolate the exact amount of energetic cost or benefit associated with that change.

TROT is designed for research and teaching rather than for operational robot work—while some 3D-printed parts break easily, they are also easy to repair and replace. Still, the results of future studies with this robot could inform commercial designs. At present, most commercial quadrupeds have fore and hind legs of the same length and style, but this test robot could reveal how to optimize the legs for the robot's intended purposes and terrains, and quantify whether the gains are worth the increase in manufacturing costs.

Researchers and enthusiasts can download the plans for the robot from U-M. The printing instructions for the parts are largely written for typical resin 3D printers, known as fused deposition modeling printers, with a stereolithography printer needed for a couple of components.

Urs is now the lead spacecraft engineer at Argo Space.

Study: The Robot of Theseus: A modular robotic testbed for legged locomotion (DOI: 10.1088/1748-3190/ae3ec1)

Journal

Bioinspiration & Biomimetics

DOI

10.1088/1748-3190/ae3ec1

Spherical assemblies of nanocrystals

Universiteit van Amsterdam

From table salt to snowflakes, and from gemstones to diamonds—we encounter crystals everywhere in daily life, usually cubic (table salt) or hexagonal (snowflakes). Researchers from Noushine Shahidzadeh's group at the UvA Institute of Physics now demonstrate how mesmerizing spherical crystal shapes arise through structures called spherulites.

A new study done in Shahidzadeh’s lab at the Institute of Physics / Van der Waals Zeeman-institute, reveals how neatly ordered (hemi-) spherical or pancake like structures in nature can emerge from completely disordered salt solutions. Moreover, scientists can now harness these structures to create advanced materials.

Structures like tiny sea urchins

Take some ordinary table salt and dissolve it in water. The salt breaks down into its tiniest constituents, ions – atoms with some electron charges added or removed. The same process can be carried out for many other materials. When the water evaporates, crystallization – the same process that forms snowflakes or rock candy – transforms these disordered mixtures of ions into elegant, structured forms: crystals. This process of crystallization underlies a wide array of techniques, from purifying medicines to crafting high performance materials such as silicon wafers that power modern technologies.

The researchers now show that for mixtures of different ions, something extraordinary occurs. Instead of forming a single flawless crystal, the matter can organize itself into so-called spherulites — mesmerizing, spherical structures that sprout like tiny sea urchins or coral heads under the microscope. They uncovered how subtle shifts in ion composition — specifically, the presence of so-called divalent ions in highly viscous mixed sulfate solutions — drive the formation of well-organized sodium sulfate nanocrystals into spherulitic shapes at room temperature.

Tess Heeremans, first author of the study and now a PhD student at AMOLF, explains how the discovery came about. “We stumbled upon the spherulites by surprise during my master’s internship with Noushine, one of those magical moments in the lab. Once we saw them evolve from our salty mixtures under the microscope, we couldn’t look away — it was so cool! I thought: Is this a crystal? It does not look like it at all! That curiosity ended up steering my master thesis in a new direction and resulted with hard work and collaboration in a publication which I am extremely proud of.”

The work shows how subtle variations in composition, viscosity and evaporation rate determine whether the crystals grow as open, spiky forms, dense spheres, or well-defined regular lattices. The work shows the amazing beauty of Nature, but the discovery also has practical applications. Beyond decoding fossil-like mineral textures once mistaken for biological remains, it offers a pathway to engineer materials with intricate internal architectures and exceptionally large surface area with designs written by the physics of nonequilibrium growth.

As Heeremans puts it: “Spherulites may look magical, but their beauty serves as record; like a snowflake shaped by the clouds it grew in, a spherulite reflects the environment of its formation. By identifying key growth conditions, we can direct crystalline salts to form these high surface-to-volume structures, which can drastically change the properties of many materials.”

Journal

Nature

DOI

10.1038/s42004-026-01892-0

Keeping an eagle eye on carbon stored in the ocean

Carbon capture and storage (CCS) is just one of many tools that the IPCC says the nations of the world have to employ to cut greenhouse gas emissions into the atmosphere. Accurate monitoring of undersea storage sites is critical to this effort

Norwegian University of Science and Technology

Modeling the Sleipner field carbon undersea storage site — image:
Kasper Hunnestad, a postdoc at the Norwegian University of Science and Technology (NTNU) Centre for Geophysical Forecasting, stands in a special tank he and colleagues have built to simulate what happens in a model of the Sleipner field CO2 storage site.
view more
Credit: Photo: Nancy Bazilchuk, NTNU

As Norway and other nations begin to scale up the storage of CO2 in undersea geologic reservoirs, research from the Norwegian University of Science and Technology (NTNU) is helping answer two important questions about this storage.

“Where has my CO2 gone? Is it leaking or not?” says Martin Landrø, an NTNU geophysicist and director of the university’s Centre for Geophysical Forecasting (CGF). “Those are the basic questions actually.”

This is like a revolution in visualization and understanding of what's happening.

Norway is home to the longest running undersea CO2 storage project in the world, at the Sleipner gas field in the North Sea. There, a total of 20 million tons of CO2 have been injected at Sleipner into a saline aquifer called the Utsira Formation.

Using a data-analysis technique called full-waveform inversion, CGF researchers have taken a closer look at data from the Sleipner site. The data are collected using geophysical methods such as seismic imaging.

A new paper from a newly minted CGF PhD, Ricardo Jose Martinez Guzman, shows how effective this technique can be in verifying where CO2 is and how much has been injected.

“Maybe 10 years ago, the full-waveform inversion from Sleipner was like wearing very foggy glasses. But this has now advanced so far that we can see all the layers and all these feeders. So, this is like a revolution in visualization and understanding of what's happening,” said Philip Ringrose, a professor in Energy Transition Geoscience at CGF.

Where is my CO2? And is it leaking?

Right now, companies use ships to tow acoustic sensors over these undersea storage formations, sailing back and forth over the formation in a grid pattern, much like you might carefully mow a lawn.

We're pushing the technology to show you can see everything with geophysics.

This takes time and money, of course. But could there be better ways to get the same information?

In areas where CO2 storage is land-based, companies can drill wells to check where the CO2 has gone, Ringrose said. But that’s not the best option in places like Norway, where storage sites can be a thousand metres or more below the seabed.

“Here we don't use wells to check where the CO2 is. We only use geophysical data. That's partly because we're offshore, but it's also because we're pushing the technology to show you can see everything with geophysics,” Ringrose said.

A tank and a several-hundred-kilo plastic model

In addition to advancing analytical tools like full-waveform inversion, CGF researchers have built a new laboratory to help better understand the complexities of undersea storage.

The new lab is centred around a 2-by-4 metre tank filled with water. Inside the tank is a several-hundred-kilo mockup of the top layer of the Utsira Formation. This is the cap rock that prevents the CO2 from leaking out of the formation

This laboratory and its big tank are the equivalent of a sandbox where CGF researchers can test different ways of measuring what’s happening in the Utsira mock-up.

And because they have 30 years of data from Sleipner, they know how the CO2 has behaved in the past and can use that for comparisons and calibration.

Challenging the system to see what happens

Kasper Hunnestad, a CGF postdoc, is responsible for the lab. He’s spent hours and hours setting up the tank with its heavy plastic model.

What we can do is to challenge the system a bit. We know what works. But what happens if you take away some of the data?

On each end of the tank are movable racks with rows of aluminium tubes, about the diameter of a garden hose and topped by black wires. Each of the wires goes to a ultrasonic sender and receiver that he can use to scan the mock-up as he changes the amount of air – his CO2 proxy – he’s injecting into the system.

The racks can be moved back and forth over the mockup at different speeds, mimicking how a conventional underwater seismic survey might be conducted.

You can see the air bubbles in the translucent plastic of the model, but more importantly, the wild tangle of wires and sensor allows him to scan the model and show exactly what has happened to his CO2 proxy over time.

“Then if you have several of these scans, you can kind of set up a timeline image of how the air has been distributed around the tank, which is the CO2 monitoring part,” he said. “That’s actually what you would do in the field.”

But the beauty of the lab is that Hunnestad can change some of the features of the system to see what happens to the information he gets.

“What we can do is to challenge the system a bit. We know what works. But what happens if you take away some of the data? What if you don’t have the luxury of having all the data, can we still see how the CO2 is distributed?” he said.

The answers could help reduce the costs of monitoring CO2 storage sites – and could help improve their accuracy.

Improving accuracy and looking ahead

CGF’s Ringrose says that CGF’s industrial partners are deeply interested in seeing results from the test laboratory.

“The competing geophysical companies who are our partners want to be in this space because they see a business opportunity. They want to be able to go to the operators and say, we can tell you where your CO2 is. It is definitely a competitive space,” he said.

Centre director Landrø thinks the future may lie in using sensing technologies such as fibre optic cables. These are the glass fibre cables that are used carry information, such as the internet and communications information across the ocean.

In other CGF research, Landrø and his colleagues have been able to use fibre optic cables offshore of Svalbard to identify and track whales. So why not CO2?

“What we foresee in the future is that if you have a storage area like this, you deploy not conventional seismic cables, but fibre optic cables and you just plough them 10 or 20 centimetres below the seabed,” he said. “That will be a challenge to do in a quick and economic way, but the fibre doesn’t cost anything.”

References:
Ricardo Martinez, Vetle Vinje, Harrison Moore, Steve Hollingworth, Philip Ringrose, Alexey Stovas; Unraveling multi-layer CO2 plumes using the entire wavefield: Case study from the Sleipner storage site. Interpretation 2025; doi: https://doi.org/10.1190/int-2025-0016

Ringrose, P., Martinez, R., Vinje, V. and Mispel, J., 2024. Estimating the Multi-Scale Distribution of Co2 Using Seismic Data at Sleipner. International Journal of Greenhouse Gas Control, Volume 151, 104581 https://doi.org/10.1016/j.ijggc.2026.104581

Ringrose, P., 2023. Storage of Carbon Dioxide in Saline Aquifers: Building confidence by forecasting and monitoring. Society of Exploration Geophysicists.
https://doi.org/10.1190/1.9781560803959

Kasper Hunnestad explains how he can move the racks with ultrasonic sensors over the mockup of the Utsira formation to simulate the way seismic surveys are conducted out in the ocean. The ultrasonic sensors are in the tall metal pipes to the left of the photo.

Credit

Photo: Nancy Bazilchuk, NTNU

The Norwegian Offshore Directorate has an atlas that lists and assesses CO2 storage capacity on the Norwegian Continental Shelf (NCS). This screenshot shows an interactive map where undersea reservoirs have been explored (the orange hatched lines) and approved (the orange crosshatched lines). The areas outlined in blue or yellow are potential storage areas. At the date of this screenshot, roughly 39.5 thousand tonnes of CO2 had been stored on the NCS. Screenshot: Norwegian Offshore Directorate

Credit

Norwegian Offshore Directorate

Journal

Interpretation

DOI

10.1190/int-2025-0016