Integrated sensing and communications (ISAC) is transforming wireless technology by combining sensing and connectivity into a single system. By leveraging existing network infrastructure, ISAC enables wireless networks to detect environmental factors, optimize radio channels, and improve energy efficiency. One key innovation is using computer vision to analyze surroundings, identifying materials like windows and concrete to refine signal transmission. This technology helps wireless devices adapt dynamically, leading to more reliable connections and smarter energy use.
ISAC enhances safety and efficiency. By using advanced sensing techniques, networks can track and differentiate between authorized and unauthorized drones, even in areas where GPS falls short. With real-world applications already taking shape, it is set to play a pivotal role in shaping the future of wireless connectivity and automation.
Steve Saunders:
Tao, what is integrated sensing and communications? I'm hearing a lot about it. Why is it important? Why is it exciting?
Tao Luo:
So, this integrated sensing communication is two parts. One is communication can help the sensing. The second one is sensing can help the communication. You can leverage existing infrastructure of the base station and be able to create a new value for wireless systems.
Steve Saunders:
Very interesting. Are there particular use cases that you can give us examples to bring this to life for our audience?
Tao Luo:
Yeah. So in this special demo we're showing here, so we're thinking of by using sensing, we're able to determine the environmental information. Then we're able to generate a radio channel. Then we can help the [inaudible 00:00:53] network device to save the energy. This is one of the applications for the sensing assisting the communications.
Steve Saunders:
Wow. Really, really cool. How does it actually work though? Talk us through that in a little bit more detail, maybe step by step.
Tao Luo:
Okay. So to make this work, we have a major two steps. So once you have a photo of an area, able to determine the majority of the material information in the environment by the computer vision. So then once the computer vision to determine, for example, there's a window in the beauty, then the next one will use a sensor to determine what type of the window you have. There's high penetration loss, medium penetration loss, or low penetration loss. With this information determined, they can know your environments much better. They can generate channel, then they help the communication systems.
Steve Saunders:
Wow. What are we looking at over here, Tao? Is this an example of how this works?
Tao Luo:
Yeah. This is the one... the first step. With the imaging information, the computer vision determines what the environmental looks like in the image environment we have.
Steve Saunders:
Okay. And can I manipulate this? Can I use it to develop... To drill down into what I want to use this for?
Tao Luo:
This one can help you [inaudible 00:02:15] for the communication.
Steve Saunders:
Right.
Tao Luo:
Because with this one, once you determine all the material, like the concrete, windows, then you can generate better radio channel by using [inaudible 00:02:28] tracing for example. Right? Once we do the better channel modeling, then it can help the communication system to be able to do better prediction or assist information-
Steve Saunders:
But I could also do manipulate? I can do what if? Like an example of, you know, if I want to change things here in terms of what I've generated and say, "Okay, but what if the physical model was changed in these ways?" Can I use this for that type of modeling?
Tao Luo:
You can also change because there's a computer vision. You know? Most of the environments don't change much, but there's always maybe a moving car or people coming around. With those one you can do the regenerative segmentation.
Steve Saunders:
Okay. Right.
Tao Luo:
More dynamic.
Steve Saunders:
Yeah.
Tao Luo:
Yeah.
Steve Saunders:
Is that a big data download? Is it a big processing task or it does it fairly easily?
Tao Luo:
It's not that complicated.
Steve Saunders:
Yeah.
Tao Luo:
Because many of the computer vision processing use well-known techniques today already. You know?
Steve Saunders:
Yeah.
Tao Luo:
Plus our own algorithm to better differential different type of the materials. Right?
Steve Saunders:
I'm hearing a lot about the low-altitude economy. I don't know what that is.
Tao Luo:
So, for example, today there's a drone flying around-
Steve Saunders:
Yeah.
Tao Luo:
... either deliver the food or deliver the goods, like Amazon. Or, in China, deliver a lunchbox, you know?
Steve Saunders:
Yeah. Sure. Yeah.
Tao Luo:
So you need to monitor this drone when it's flying around. Traditional signal image processing may not be sufficient sometimes, right? Which is a sensing. Can determine if it's an authorized drone or unauthorized drone, where they are and tracking them.
Steve Saunders:
Yeah.
Tao Luo:
You can more safely operate this drone business.
Steve Saunders:
Well that sounds pretty important. Would I combine that with some sort of three-dimensional GPS system?
Tao Luo:
You can combine it. All is complementary to each other.
Steve Saunders:
All complementary.
Tao Luo:
Yeah.
Steve Saunders:
Yeah.
Tao Luo:
GPS may not work well when you have a high-rise building.
Steve Saunders:
Yeah.
Tao Luo:
But this sensing can work still well in those scenarios.
Steve Saunders:
Yeah. So you need both?
Tao Luo:
Yeah, I would say they will complement each other.
Steve Saunders:
So in China, I know that you already have packages delivered by drone and you also have autonomous vehicles going around and robotics and automation and predictive analytics. It feels like China's about 10 years ahead of the rest of the world. Do you think that's accurate?
Tao Luo:
I would say there's a more need there because it delivers the goods and also avoids the traffic and also the deliver lunchbox, which is time-critical for the people. Right?
Steve Saunders:
Yeah. Very.
Tao Luo:
So I see the use case is pretty useful to help the economy over there.
Steve Saunders:
Yeah. Yeah, it's fantastic.
Tao Luo:
Yeah.
Steve Saunders:
This is really cool.
Tao Luo:
Yeah.
Steve Saunders:
This is all your own work?
Tao Luo:
Yeah. This is our own work. Yes.
Steve Saunders:
That's fantastic.
Tao Luo:
Right. Yeah.
Steve Saunders:
Yeah. Congratulations.
Tao Luo:
Thank you.
Steve Saunders:
Really appreciate it. Thank you for giving us the tour, Tao.
Tao Luo:
Right. Thank you so much. Okay.
Steve Saunders:
Thank you.