0:02 to in the conversations I've had during
0:03 Mobile World Congress around 6G one
0:05 thing keeps coming up and that's
0:07 integrated sensing and communication so
0:09 to start maybe you could just sort of
0:12 explain to us what that is so that's a
0:14 good question so integrated sensing and
0:16 communication which means you can
0:19 leverage existing Cellar infrastructure
0:21 for the sensing and communication
0:24 traditional c s only for communication
0:26 now you add additional function for
0:28 Network that's why it's called integrate
0:31 sensing and Communications okay so it
0:33 seems like there'd be two sort of
0:35 primary buckets of use cases for this
0:37 one the network understands its own
0:39 physical environment so it become more
0:41 adaptable as things around it change but
0:43 then you could also use that sensing
0:46 capability to do other types of services
0:48 that might be of use to various vertical
0:49 Industries but can you kind of take us
0:52 through both of those yeah so first you
0:54 can do the sensing assisity
0:56 communication you know you can improve
0:58 the radio performance by knowing the
1:00 environments better you you know for
1:02 example the demo we showing here you
1:04 know they were showing by using this
1:07 understand this environments better you
1:09 can do better communication on the other
1:11 hand you also have other use case for
1:13 example security detect a person you
1:16 know or drone detection flying around
1:20 you can track detect so you can better
1:23 uh track the uh enable is a l attitude
1:25 economy for example in different
1:27 countries you know so on what sort of
1:29 timeline should we think about this
1:31 going from uh research Focus to
1:33 something that would be a little more
1:35 real world a little more tangible so in
1:37 some of the case it's already happening
1:39 in some of the uh country you know they
1:42 doing the trial for this Jun you know on
1:44 the other hand there also are there more
1:47 use case weely more research so it takes
1:49 some time to actually to realize the
1:52 full potential of this one then I wanted
1:54 to also get some uh perspective from you
1:58 on a wireless AI so I know this is a
1:59 priority for you and your colleagues can
2:01 you give us a bit of an overview of uh
2:04 where your focus today is today our
2:06 Focus AI is one to improve the wireless
2:09 air interface which means you can do
2:12 better algorithm save the energy either
2:14 at the LK side or device side you know
2:16 and also potentially even improve the
2:18 performance so it's focus on the air
2:21 interface the one we're currently doing
2:24 and so then how do you operationalize
2:27 Wireless AI I guess if you have a a
2:30 model of a of a real Network and all of
2:32 the variability that that entails you
2:34 know how do you how do you take that and
2:36 then create that sort of closed loop
2:39 where you're able to reap the benefit of
2:41 the adaptability that you put into the
2:44 network yeah there's a two aspect on
2:46 this one one is called a model switching
2:49 so want you move to one scenario to
2:50 another scenario for example Urban to
2:52 Suburban you may want to do the model
2:54 switching you have a different model has
2:57 been trained know on the other hand also
2:59 want to say is UND devised training you
3:02 continue just learning just like a human
3:04 you get more information you update your
3:06 model and use it
3:08 immediately so is it fair to think of it
3:10 as a device to network Continuum where
3:12 there's intelligence throughout and it's
3:15 uh continuing to learn continuing to
3:18 adapt and improve yes that would be best
3:21 optimize your performance and also
3:23 provide a lot of advantage to able to
3:26 scale this AI AML model for air
3:28 interface right very good well T I
3:29 really appreciate you taking a few
3:31 minutes to tell us about the work you
3:32 and your colleagues are doing thank you
3:34 so much [Music]
Chrome Extension