WEBVTT 00:00:12.080 --> 00:00:14.150 What's up everybody? Koiwire here with a 00:00:14.160 --> 00:00:16.550 special edition of CNN 10. Today we're 00:00:16.560 --> 00:00:18.150 taking a deep dive into the world of 00:00:18.160 --> 00:00:20.550 artificial intelligence and self-driving 00:00:20.560 --> 00:00:23.189 cars. Today we're going to Curiosity Lab 00:00:23.199 --> 00:00:25.109 in Peach Tree Corners, Georgia to 00:00:25.119 --> 00:00:27.349 experience this tech firsthand. This is 00:00:27.359 --> 00:00:31.429 my first time in a Whimo. Let's go. 00:00:31.439 --> 00:00:32.709 You could call this one of the most 00:00:32.719 --> 00:00:35.270 advanced smart cities in the world. 00:00:35.280 --> 00:00:37.750 Tomorrow's technologies being tested and 00:00:37.760 --> 00:00:40.389 created today. The color printer was 00:00:40.399 --> 00:00:42.389 invented here in Technology Park. The 00:00:42.399 --> 00:00:45.590 Haze modem, too. Now, autonomous or 00:00:45.600 --> 00:00:48.069 self-driving vehicles are being steered 00:00:48.079 --> 00:00:51.190 to new frontiers. Air package delivery 00:00:51.200 --> 00:00:54.069 systems via drones. Mobile delivery 00:00:54.079 --> 00:00:57.430 robots tested and perfected. Where are 00:00:57.440 --> 00:00:59.029 we? What is this place? I feel like I 00:00:59.039 --> 00:01:00.150 stepped into the future. 00:01:00.160 --> 00:01:02.229 >> This is Curiosity Lab. We are a 00:01:02.239 --> 00:01:04.070 nonprofit just inside the city of Peach 00:01:04.080 --> 00:01:06.230 Tree Corners. And what we do is we help 00:01:06.240 --> 00:01:08.070 companies test and deploy their 00:01:08.080 --> 00:01:10.630 technology on public infrastructure. So 00:01:10.640 --> 00:01:12.789 that meant always sounds super cool, but 00:01:12.799 --> 00:01:14.390 it is. I mean, think about all the stuff 00:01:14.400 --> 00:01:15.990 that public infrastructure entails, 00:01:16.000 --> 00:01:18.230 right? Roadways, intersections, even 00:01:18.240 --> 00:01:19.910 airspace. Um, a lot of different 00:01:19.920 --> 00:01:21.109 technologies work on those things. 00:01:21.119 --> 00:01:22.310 >> And that's very important because we're 00:01:22.320 --> 00:01:23.830 living at a time where we're seeing more 00:01:23.840 --> 00:01:25.670 and more of it, hearing more and more 00:01:25.680 --> 00:01:28.870 about it. They might be taxi cabs that 00:01:28.880 --> 00:01:31.270 are flying through this city. So how 00:01:31.280 --> 00:01:33.109 well all do you help companies prepare 00:01:33.119 --> 00:01:34.550 us for that type of future? 00:01:34.560 --> 00:01:37.109 >> Sure. So when it comes to an environment 00:01:37.119 --> 00:01:38.789 like what is the ideal environment for 00:01:38.799 --> 00:01:40.789 testing? So you always start in a closed 00:01:40.799 --> 00:01:42.069 environment, right? There's there's no 00:01:42.079 --> 00:01:43.670 things coming in in terms of variables. 00:01:43.680 --> 00:01:45.190 You want to make sure the product works. 00:01:45.200 --> 00:01:46.950 So in the case of a vehicle on a 00:01:46.960 --> 00:01:48.230 roadway, that would be like a closed 00:01:48.240 --> 00:01:50.710 track. But we also know that from there, 00:01:50.720 --> 00:01:51.910 how do you get into an urban 00:01:51.920 --> 00:01:53.749 environment? How do you get into a real 00:01:53.759 --> 00:01:55.190 city? And you've got to have that middle 00:01:55.200 --> 00:01:57.590 ground. We have a closed three mile loop 00:01:57.600 --> 00:02:00.149 inside a 500 acre ecosystem so that 00:02:00.159 --> 00:02:02.789 cars, technology, drones, whatever it 00:02:02.799 --> 00:02:04.550 is, can test in a real environment where 00:02:04.560 --> 00:02:06.469 there's real pedestrians, real cars, 00:02:06.479 --> 00:02:08.710 real people, real intersections. So you 00:02:08.720 --> 00:02:10.309 can start to get the data and learn how 00:02:10.319 --> 00:02:12.150 these technologies work in real time 00:02:12.160 --> 00:02:13.830 with real elements so that you know, 00:02:13.840 --> 00:02:15.830 hey, we're ready to get into the city. 00:02:15.840 --> 00:02:18.949 >> I just saw a robot fly by. I don't know 00:02:18.959 --> 00:02:19.750 what it was, but 00:02:19.760 --> 00:02:20.949 >> oh my gosh, I think it might have been 00:02:20.959 --> 00:02:22.869 Cheetah. She's a cute little thing. 00:02:22.879 --> 00:02:24.309 Basically, she's a helping hand 00:02:24.319 --> 00:02:26.309 essentially. She is a a robot that 00:02:26.319 --> 00:02:27.830 follows a subject in front of it. So, 00:02:27.840 --> 00:02:29.510 it's not a self-driving one. It's not 00:02:29.520 --> 00:02:31.190 like it's doing delivery, but it's more 00:02:31.200 --> 00:02:32.630 like, you know, a robotic dog that 00:02:32.640 --> 00:02:33.350 follows you everywhere. 00:02:33.360 --> 00:02:34.390 >> You don't need to carry a backpack 00:02:34.400 --> 00:02:34.869 anymore. 00:02:34.879 --> 00:02:36.309 >> It's exactly what it is. It's a rolling 00:02:36.319 --> 00:02:39.030 robotic backpack. You open up the uh vat 00:02:39.040 --> 00:02:40.390 inside of it, and you know, they have 00:02:40.400 --> 00:02:41.589 different inserts. I mean, you can make 00:02:41.599 --> 00:02:43.910 it follow you like a tailgate. Um, but, 00:02:43.920 --> 00:02:44.949 you know, it also increases 00:02:44.959 --> 00:02:47.589 accessibility for things like disabled 00:02:47.599 --> 00:02:49.190 folk or even elderly and, you know, 00:02:49.200 --> 00:02:50.710 walkable communities if you're at the 00:02:50.720 --> 00:02:52.229 airport and you need help carrying 00:02:52.239 --> 00:02:54.790 stuff. Um, so it's cool. The fun fact 00:02:54.800 --> 00:02:57.190 there is that its design was based off 00:02:57.200 --> 00:02:59.430 of the droids in Star Wars. So they're 00:02:59.440 --> 00:03:01.430 they're quite cute, I think. 00:03:01.440 --> 00:03:03.990 >> Now you're speaking my language. Where's 00:03:04.000 --> 00:03:06.229 my lightsaber? I feel cars whizzing 00:03:06.239 --> 00:03:08.229 behind me. So tell me what's happening 00:03:08.239 --> 00:03:09.670 here. What are we seeing? What are we 00:03:09.680 --> 00:03:11.110 learning from what we're seeing? 00:03:11.120 --> 00:03:13.350 >> Yeah, absolutely. So we talk about, you 00:03:13.360 --> 00:03:16.630 know, the roadway. Um that is a very 00:03:16.640 --> 00:03:18.229 loaded thing to think about. There's so 00:03:18.239 --> 00:03:19.670 many things going on at all different 00:03:19.680 --> 00:03:21.670 times on the roadway. And as drivers, 00:03:21.680 --> 00:03:23.350 you know, we learn to get used to these 00:03:23.360 --> 00:03:24.790 things. You learn when someone's going 00:03:24.800 --> 00:03:26.710 to cut in front of you. You learn how to 00:03:26.720 --> 00:03:28.149 interact with different lights and 00:03:28.159 --> 00:03:29.509 different signals and different signs. 00:03:29.519 --> 00:03:30.789 You know, what we're doing here is we're 00:03:30.799 --> 00:03:32.710 using uh cameras and we're putting 00:03:32.720 --> 00:03:34.550 analytics over those cameras to show you 00:03:34.560 --> 00:03:36.309 what's happening in here. So, um this 00:03:36.319 --> 00:03:38.470 can be a forum of AI, right? So, it 00:03:38.480 --> 00:03:39.750 could be doing things like object 00:03:39.760 --> 00:03:41.030 detection. As you can see here, it's 00:03:41.040 --> 00:03:42.949 looking at vehicle vehicle. It could 00:03:42.959 --> 00:03:45.350 clock a, you know, a bicyclist or a 00:03:45.360 --> 00:03:46.789 pedestrian, which we're seeing a couple 00:03:46.799 --> 00:03:49.350 pedestrians come right there. So, it 00:03:49.360 --> 00:03:51.110 says unknown and that's a part of 00:03:51.120 --> 00:03:52.630 testing, right? Is that, you know, it's 00:03:52.640 --> 00:03:54.229 still learning. It says person now as 00:03:54.239 --> 00:03:56.070 it's getting closer to the camera. 00:03:56.080 --> 00:03:57.429 >> Um, but that's part of it, right? Is 00:03:57.439 --> 00:03:58.630 there are times where a truck comes 00:03:58.640 --> 00:04:00.149 through and it and you know, it might 00:04:00.159 --> 00:04:01.750 say person and you're like, well, that's 00:04:01.760 --> 00:04:03.270 definitely not a person. But that's why 00:04:03.280 --> 00:04:04.710 you need the testing environment so they 00:04:04.720 --> 00:04:06.869 can train their AI models to do better. 00:04:06.879 --> 00:04:08.949 >> Okay. So, we've spoken about the cities, 00:04:08.959 --> 00:04:10.949 the roadways, the intersections of 00:04:10.959 --> 00:04:12.550 tomorrow. 00:04:12.560 --> 00:04:14.550 Also, there will be the cars of tomorrow 00:04:14.560 --> 00:04:16.390 that work better with that. And I 00:04:16.400 --> 00:04:18.229 understand there's a garage where we can 00:04:18.239 --> 00:04:19.830 maybe check out one of these cars. Put 00:04:19.840 --> 00:04:20.629 me to work. 00:04:20.639 --> 00:04:22.469 >> All right, let's do it. Let's go. 00:04:22.479 --> 00:04:24.390 >> The company testing out their fleet of 00:04:24.400 --> 00:04:26.469 autonomous vehicles at Curiosity Lab 00:04:26.479 --> 00:04:29.189 during our visit is called May Mobility. 00:04:29.199 --> 00:04:31.430 These cars use similar technology to the 00:04:31.440 --> 00:04:33.110 brands we may be familiar with like 00:04:33.120 --> 00:04:35.909 Whimo. But this company is particularly 00:04:35.919 --> 00:04:38.550 interested in making AVs that are more 00:04:38.560 --> 00:04:40.390 accessible for passengers who use 00:04:40.400 --> 00:04:42.390 wheelchairs, scooters, or need other 00:04:42.400 --> 00:04:43.430 accommodations. 00:04:43.440 --> 00:04:45.510 >> In order for an autonomous vehicle to 00:04:45.520 --> 00:04:47.189 work, you have to have a bunch of 00:04:47.199 --> 00:04:49.030 different technology, right? So, you've 00:04:49.040 --> 00:04:51.590 got to have the hardware, and I'll walk 00:04:51.600 --> 00:04:52.790 through that in a minute. And you have 00:04:52.800 --> 00:04:54.150 to have the software, which is pretty 00:04:54.160 --> 00:04:56.230 much the brains of the technology. So 00:04:56.240 --> 00:04:58.790 the AI component, you're using 00:04:58.800 --> 00:05:01.189 human-like reasoning, which is basically 00:05:01.199 --> 00:05:03.189 the software component, and combining it 00:05:03.199 --> 00:05:04.710 with all the things you're gathering 00:05:04.720 --> 00:05:07.430 from the hardware so that the technology 00:05:07.440 --> 00:05:09.350 can make thousands and thousands and 00:05:09.360 --> 00:05:11.830 thousands of scenarios and decisions in 00:05:11.840 --> 00:05:13.029 real time. 00:05:13.039 --> 00:05:15.029 >> So you'll see here different 00:05:15.039 --> 00:05:16.870 technologies. A lot of people say this 00:05:16.880 --> 00:05:20.150 looks like a cup holder. It is not. It's 00:05:20.160 --> 00:05:22.150 actually technology. 00:05:22.160 --> 00:05:24.870 And you'll see cameras here and LAR 00:05:24.880 --> 00:05:27.350 sensors. We have a total of nine cameras 00:05:27.360 --> 00:05:29.590 all around the vehicle. And then we have 00:05:29.600 --> 00:05:32.310 five LARs and radars. The other big guy 00:05:32.320 --> 00:05:35.430 here, the top hat, is basically um our 00:05:35.440 --> 00:05:38.710 largest LAR. And that is able to look 00:05:38.720 --> 00:05:41.029 out further away so that we can see 00:05:41.039 --> 00:05:43.909 what's happening in real time. 00:05:43.919 --> 00:05:46.150 >> While autonomous vehicles are relatively 00:05:46.160 --> 00:05:48.629 new, lidar has actually been around for 00:05:48.639 --> 00:05:50.710 quite some time. It was developed in the 00:05:50.720 --> 00:05:53.110 1960s and it was actually used on the 00:05:53.120 --> 00:05:55.749 Apollo space missions to map out the 00:05:55.759 --> 00:05:59.029 surface of the moon. It uses lasers, 00:05:59.039 --> 00:06:00.950 shoots them out everywhere to measure 00:06:00.960 --> 00:06:02.790 distances and create an incredibly 00:06:02.800 --> 00:06:05.110 accurate picture of an environment. 00:06:05.120 --> 00:06:07.270 Here's how it works. The system fires 00:06:07.280 --> 00:06:10.070 millions of laser pulses 00:06:10.080 --> 00:06:12.309 every second in every direction. They 00:06:12.319 --> 00:06:14.230 bounce off surrounding objects to create 00:06:14.240 --> 00:06:16.710 a detailed 3D picture of its 00:06:16.720 --> 00:06:18.710 surroundings. That includes everything 00:06:18.720 --> 00:06:20.710 from buildings to other cars, people, 00:06:20.720 --> 00:06:23.189 and animals. It's similar in a way to 00:06:23.199 --> 00:06:25.430 echolocation, the system used by bats, 00:06:25.440 --> 00:06:27.749 whales, and dolphins to navigate and to 00:06:27.759 --> 00:06:30.309 hunt prey. But this system swaps the 00:06:30.319 --> 00:06:32.550 sound waves for light. You may have seen 00:06:32.560 --> 00:06:34.950 the rapidly spinning mechanisms on some 00:06:34.960 --> 00:06:36.790 of these autonomous cars. Well, that 00:06:36.800 --> 00:06:39.990 allows them to have a 360° field of view 00:06:40.000 --> 00:06:42.629 to help eliminate blind spots and safely 00:06:42.639 --> 00:06:45.749 navigate the world around them. 00:06:45.759 --> 00:06:47.990 Pop quiz hot shot. Self-driving cars 00:06:48.000 --> 00:06:50.150 often use lidar to scan surroundings. 00:06:50.160 --> 00:06:51.909 What does lidar stand for? Light 00:06:51.919 --> 00:06:53.590 detection and ranging. Laser 00:06:53.600 --> 00:06:55.430 identification and radar. Light 00:06:55.440 --> 00:06:57.270 identification and response. Looking 00:06:57.280 --> 00:07:00.870 into dangers around the route. Answer is 00:07:00.880 --> 00:07:03.510 light detection and ranging. Lidar is a 00:07:03.520 --> 00:07:05.270 remote sensing method used to examine 00:07:05.280 --> 00:07:07.990 the surface of the earth. So what you 00:07:08.000 --> 00:07:10.150 want to do is you want to use a 00:07:10.160 --> 00:07:12.870 combination of historical data, right, 00:07:12.880 --> 00:07:15.670 which it's accumulated over time based 00:07:15.680 --> 00:07:17.990 on different locations. So for example, 00:07:18.000 --> 00:07:20.950 let's say you're in Georgia, hot 00:07:20.960 --> 00:07:24.070 weather, right? You're in Phoenix, 00:07:24.080 --> 00:07:27.270 Arizona, hot but dry weather, and you're 00:07:27.280 --> 00:07:30.790 in Michigan where you have snow. So, you 00:07:30.800 --> 00:07:32.390 want to be able to take all of those 00:07:32.400 --> 00:07:34.710 elements and have the vehicle learn from 00:07:34.720 --> 00:07:36.870 all the different weather conditions so 00:07:36.880 --> 00:07:38.710 that when it's a location, if we happen 00:07:38.720 --> 00:07:41.110 to get snow in Georgia this year, the 00:07:41.120 --> 00:07:43.189 vehicle will know what to do. 00:07:43.199 --> 00:07:45.670 >> But I think about, you know, my 00:07:45.680 --> 00:07:47.990 daughters when they're old enough to be 00:07:48.000 --> 00:07:50.070 going places, 00:07:50.080 --> 00:07:51.749 I think that I would feel comfortable 00:07:51.759 --> 00:07:54.070 with them being in a vehicle all to 00:07:54.080 --> 00:07:56.309 themselves. The tech is only going to 00:07:56.319 --> 00:07:58.629 get better, right? And you would think 00:07:58.639 --> 00:08:00.710 as there are more and more autonomous 00:08:00.720 --> 00:08:03.510 vehicles on the roadways that are now 00:08:03.520 --> 00:08:05.430 communicating with each other, they're 00:08:05.440 --> 00:08:07.909 like computer brains communicating with 00:08:07.919 --> 00:08:10.390 other drivers. And so you would think 00:08:10.400 --> 00:08:13.189 that would be even safer than kind of 00:08:13.199 --> 00:08:15.189 guessing and wondering what a human 00:08:15.199 --> 00:08:16.390 might do. 00:08:16.400 --> 00:08:19.270 >> There's a lot of um transparency and 00:08:19.280 --> 00:08:21.430 visibility into a vehicle. you have 00:08:21.440 --> 00:08:25.749 cameras everywhere and um so you you can 00:08:25.759 --> 00:08:28.230 see what's happening and so for them in 00:08:28.240 --> 00:08:29.990 a lot of ways they're probably a lot 00:08:30.000 --> 00:08:32.389 safer you know in the vehicle we all 00:08:32.399 --> 00:08:35.190 know that you know last year I think 00:08:35.200 --> 00:08:38.230 there were over 40,000 deaths in the US 00:08:38.240 --> 00:08:41.029 from um accidents road accidents that's 00:08:41.039 --> 00:08:43.269 things like you know falling asleep 00:08:43.279 --> 00:08:46.870 texting and driving intoxication so all 00:08:46.880 --> 00:08:49.350 of those things can happen you know with 00:08:49.360 --> 00:08:51.910 with your child, right, for the future. 00:08:51.920 --> 00:08:54.470 So, and I would argue that for the 00:08:54.480 --> 00:08:56.710 future and maybe even today that these 00:08:56.720 --> 00:08:58.470 vehicles are, you know, safe and 00:08:58.480 --> 00:09:00.710 hopefully that's their future. And then 00:09:00.720 --> 00:09:02.949 you've got to think too, there's others 00:09:02.959 --> 00:09:05.110 as well. I mean, you have kids, right? 00:09:05.120 --> 00:09:07.670 But there's um people with disabilities, 00:09:07.680 --> 00:09:09.910 great for them. And then, you know, 00:09:09.920 --> 00:09:11.910 people who are elderly, right? They can 00:09:11.920 --> 00:09:14.630 come in and use those vehicles. In um 00:09:14.640 --> 00:09:17.590 Minnesota, we have um one of our biggest 00:09:17.600 --> 00:09:20.949 groups of riders are seniors and the 00:09:20.959 --> 00:09:23.110 middle school kids because they 00:09:23.120 --> 00:09:24.310 >> not seniors in high school. 00:09:24.320 --> 00:09:26.630 >> No, seniors. Seniors and middle 00:09:26.640 --> 00:09:28.949 schoolers and middle schoolers because 00:09:28.959 --> 00:09:30.949 you know a lot of times they don't have 00:09:30.959 --> 00:09:33.829 a way to get home um you know after 00:09:33.839 --> 00:09:36.630 after school activities and so this is a 00:09:36.640 --> 00:09:38.710 way for them to do it safely. 00:09:38.720 --> 00:09:40.230 >> What do you think? I want you to press 00:09:40.240 --> 00:09:41.750 pause for a moment. Discuss with your 00:09:41.760 --> 00:09:43.990 friends if a self-driving car service 00:09:44.000 --> 00:09:45.910 was widely available in your city or 00:09:45.920 --> 00:09:47.829 town. Maybe they already are. Would you 00:09:47.839 --> 00:09:49.670 use them? Would you trust AI over 00:09:49.680 --> 00:09:51.269 getting your own driver's license? What 00:09:51.279 --> 00:09:53.430 are the risks? What are the benefits? 00:09:53.440 --> 00:09:54.790 Thanks for taking a ride with me today 00:09:54.800 --> 00:09:56.949 on this special edition of CNN 10. I 00:09:56.959 --> 00:09:58.630 learned a lot today. I hope you did, 00:09:58.640 --> 00:10:03.959 too. Be kind, stay curious, and rise up.