Let's face it: Driving is dangerous. It's one of the things that we don't like to think about, but the fact that religious icons and good luck charms show up on dashboards around the world betrays the fact that we know this to be true. Car accidents are the leading cause of death in people ages 16 to 19 in the United States -- leading cause of death -- and 75 percent of these accidents have nothing to do with drugs or alcohol.
Suočimo se s tim: vožnja automobila je opasna. To je jedna od stvari o kojima ne volimo da mislimo, ali činjenica da se religiozne ikone i amajlije pojavljuju na instrument-tablama širom sveta odaje činjenicu da znamo da je ovo istina. Saobraćajne nesreće su vodeći uzrok smrti kod ljudi između 16 i 19 godina u SAD - vodeći uzrok smrti - i 75% ovih nesreća nema nikakve veze sa drogama ili alkoholom.
So what happens? No one can say for sure, but I remember my first accident. I was a young driver out on the highway, and the car in front of me, I saw the brake lights go on. I'm like, "Okay, all right, this guy is slowing down, I'll slow down too." I step on the brake. But no, this guy isn't slowing down. This guy is stopping, dead stop, dead stop on the highway. It was just going 65 -- to zero? I slammed on the brakes. I felt the ABS kick in, and the car is still going, and it's not going to stop, and I know it's not going to stop, and the air bag deploys, the car is totaled, and fortunately, no one was hurt. But I had no idea that car was stopping, and I think we can do a lot better than that. I think we can transform the driving experience by letting our cars talk to each other.
Pa šta se dešava? Niko ne zna tačno, ali sećam se svoje prve nesreće. Bila sam mladi vozač na autoputu, i videla sam da su se upalila stop svetla vozila ispred mene. Rekla sam: "Okej, sve je u redu, ovaj tip usporava, i ja ću da usporim." Pritisnula sam kočnicu. Ali ne, ovaj tip ne usporava. On staje, u mestu, na autoputu. Išao je 110 na sat - do nule? Nagazila sam kočnicu. Osetila sam kako se aktivira ABS, i auto još uvek ide, i neće se zaustaviti i znam da se neće zaustaviti, i aktivira se vazdušni jastuk, auto je uništen, i na sreću, niko nije povređen. Ali nisam imala pojma da će taj auto stati, i mislim da to možemo da radimo puno bolje. Mislim da možemo preobratiti doživljaj vožnje tako što ćemo automobilima dozvoliti da međusobno pričaju.
I just want you to think a little bit about what the experience of driving is like now. Get into your car. Close the door. You're in a glass bubble. You can't really directly sense the world around you. You're in this extended body. You're tasked with navigating it down partially-seen roadways, in and amongst other metal giants, at super-human speeds. Okay? And all you have to guide you are your two eyes. Okay, so that's all you have, eyes that weren't really designed for this task, but then people ask you to do things like, you want to make a lane change, what's the first thing they ask you do? Take your eyes off the road. That's right. Stop looking where you're going, turn, check your blind spot, and drive down the road without looking where you're going. You and everyone else. This is the safe way to drive. Why do we do this? Because we have to, we have to make a choice, do I look here or do I look here? What's more important? And usually we do a fantastic job picking and choosing what we attend to on the road. But occasionally we miss something. Occasionally we sense something wrong or too late. In countless accidents, the driver says, "I didn't see it coming." And I believe that. I believe that. We can only watch so much.
Želim da na trenutak razmislite o tome kakav je sada doživljaj vožnje. Uđete u automobil. Zatvorite vrata. U staklenom ste zvonu. Ne možete direktno osetiti svet oko vas. U produženom ste telu. Imate zadatak da njime idete putevima koje vidite delimično, među drugim metalnim divovima, pri nadljudskim brzinama. U redu? Sve što vas vodi su vaša dva oka. To je sve što imate, oči koje nisu baš stvorene za ovaj zadatak, ali vas onda ljudi pitaju da radite stvari poput menjanja traka na putu, šta je prva stvar koju traže od vas? Da sklonite oči s puta. Tako je. Prestanete da gledate kuda idete, skrenete, proverite mrtvi ugao, i vozite putem bez gledanja kuda idete. Vi i svi ostali. Ovo je bezbedan način vožnje. Zašto radimo ovo? Jer moramo, moramo da odaberemo, da li da gledam ovde ili onde? Šta je bitnije? Obično fantastično odaberemo to čemu ćemo posvetiti pažnju na putu. Ali povremeno nam nešto izmakne. Povremeno nešto opazimo na pogrešan način ili prekasno. U velikom broju nesreća, vozači kažu: "Nisam video da dolazi." I ja verujem u to. Verujem u to. Možemo videti samo određeni deo toga.
But the technology exists now that can help us improve that. In the future, with cars exchanging data with each other, we will be able to see not just three cars ahead and three cars behind, to the right and left, all at the same time, bird's eye view, we will actually be able to see into those cars. We will be able to see the velocity of the car in front of us, to see how fast that guy's going or stopping. If that guy's going down to zero, I'll know.
Ali sada postoji tehnologija koja može da nam pomogne da to unapredimo. U budućnosti će automobili međusobno razmenjivati informacije, i moći ćemo da vidimo ne samo ispred tri automobila i iza tri automobila, levo i desno, i sve u isto vreme, ptičju perspektivu, moći ćemo da vidimo i unutar tih automobila. Moći ćemo da vidimo brzinu automobila ispred nas, da vidimo koliko brzo ide ili se zaustavlja. Ako će skroz stati, ja ću to da znam.
And with computation and algorithms and predictive models, we will be able to see the future. You may think that's impossible. How can you predict the future? That's really hard. Actually, no. With cars, it's not impossible. Cars are three-dimensional objects that have a fixed position and velocity. They travel down roads. Often they travel on pre-published routes. It's really not that hard to make reasonable predictions about where a car's going to be in the near future. Even if, when you're in your car and some motorcyclist comes -- bshoom! -- 85 miles an hour down, lane-splitting -- I know you've had this experience -- that guy didn't "just come out of nowhere." That guy's been on the road probably for the last half hour. (Laughter) Right? I mean, somebody's seen him. Ten, 20, 30 miles back, someone's seen that guy, and as soon as one car sees that guy and puts him on the map, he's on the map -- position, velocity, good estimate he'll continue going 85 miles an hour. You'll know, because your car will know, because that other car will have whispered something in his ear, like, "By the way, five minutes, motorcyclist, watch out." You can make reasonable predictions about how cars behave. I mean, they're Newtonian objects. That's very nice about them.
S proračunima, algoritmima i modelima predviđanja moći ćemo da vidimo budućnost. Možda mislite da je to nemoguće. Kako predvideti budućnost? To je jako teško. Zapravo nije. S automobilima, nije nemoguće. Automobili su trodimenzionalni objekti s fiksiranom pozicijom i brzinom. Kreću se putevima. Često unapred poznatim trasama. Zaista nije teško napraviti razumna predviđanja o tome gde će automobil biti u bliskoj budućnosti. Čak i ako ste u svojim kolima i neki motociklista prođe - vrum! - 135 kilometara na sat, menjajući trake - znam da ste iskusili ovo - taj tip se nije samo "pojavio niotkuda." Taj tip je verovatno bio na putu poslednjih pola sata. (Smeh) Zar ne? Mislim, neko ga je video. Pre nekih 30 - 50 kilometara, neko ga je video, i čim ga vidi jedan automobil i stavi ga na mapu, on je na mapi - pozicija, brzina, dobra procena da će nastaviti da ide 135km/h. Vi ćete to znati, jer će vaš automobil to znati, jer će mu to šapnuti neki drugi automobil: "E da, za pet minuta, motociklista, pazi se." Možete imati razumna predviđanja o tome kako će se ponašati automobili. To su Njutnovski objekti. To je lepa stvar u vezi sa njima.
So how do we get there? We can start with something as simple as sharing our position data between cars, just sharing GPS. If I have a GPS and a camera in my car, I have a pretty precise idea of where I am and how fast I'm going. With computer vision, I can estimate where the cars around me are, sort of, and where they're going. And same with the other cars. They can have a precise idea of where they are, and sort of a vague idea of where the other cars are. What happens if two cars share that data, if they talk to each other? I can tell you exactly what happens. Both models improve. Everybody wins. Professor Bob Wang and his team have done computer simulations of what happens when fuzzy estimates combine, even in light traffic, when cars just share GPS data, and we've moved this research out of the computer simulation and into robot test beds that have the actual sensors that are in cars now on these robots: stereo cameras, GPS, and the two-dimensional laser range finders that are common in backup systems. We also attach a discrete short-range communication radio, and the robots talk to each other. When these robots come at each other, they track each other's position precisely, and they can avoid each other.
Kako doći do toga? Možemo početi s nečim tako jednostavnim poput razmenjivanja podataka o poziciji među automobilima, samo razmenom GPS-a. Ako u kolima imam GPS i kameru, imam prilično dobar osećaj toga gde sam i koliko brzo se krećem. Sa kompjuterskim vidom, mogu da procenim gde se nalaze kola oko mene, na neki način i kuda idu. Isto je sa drugim automobilima. Mogu da imaju precizan osećaj o tome gde su, i nejasan osećaj o tome gde su drugi automobili. Šta se desi ako dva automobila dele te podatke, ako razgovaraju jedan s drugim? Reći ću vam tačno šta se dešava. Oba modela se poboljšaju. Svi su na dobitku. Profesor Bob Veng i njegov tim su uradili kompjuterske simulacije toga šta se dešava kada se nejasne procene kombinuju, čak i u lakšem saobraćaju kada automobili samo dele GPS podatke, i ovo istraživanje smo prebacili iz kompjuterske simulacije u robote za testiranje koji imaju prave senzore koji su sada u automobilima na ovim robotima: stereo kamere, GPS, i dvodimenzionalne laserske detektore dometa koji su česti u sistemima za podršku. Takođe stavljamo diskretni kratkodometni radio za komunikaciju i roboti međusobno pričaju. Kada se ovi roboti susretnu, oni jedan drugom precizno prate poziciju i mogu da se mimoiđu.
We're now adding more and more robots into the mix, and we encountered some problems. One of the problems, when you get too much chatter, it's hard to process all the packets, so you have to prioritize, and that's where the predictive model helps you. If your robot cars are all tracking the predicted trajectories, you don't pay as much attention to those packets. You prioritize the one guy who seems to be going a little off course. That guy could be a problem. And you can predict the new trajectory. So you don't only know that he's going off course, you know how. And you know which drivers you need to alert to get out of the way.
Sada u priču dodajemo sve više i više robota i naišli smo na neke probleme. Jedan od problema je, kada dođe do previše čavrljanja, teško je obraditi sve podatke, tako da morate da ih poređate po prioritetu i tu vam pomaže model predviđanja. Ako svi vaši robotski automobili prate predviđene putanje na te podatke ne obraćate toliko pažnje. Prioritet date onom tipu koji izgleda kao da ide malo van putanje. Taj tip bi mogao da bude problematičan. I možete predvideti novu putanju. Sada nećete znati samo da ide van putanje, nego i kako to radi. I znate koje vozače morate obavestiti da se sklone s puta.
And we wanted to do -- how can we best alert everyone? How can these cars whisper, "You need to get out of the way?" Well, it depends on two things: one, the ability of the car, and second the ability of the driver. If one guy has a really great car, but they're on their phone or, you know, doing something, they're not probably in the best position to react in an emergency. So we started a separate line of research doing driver state modeling. And now, using a series of three cameras, we can detect if a driver is looking forward, looking away, looking down, on the phone, or having a cup of coffee. We can predict the accident and we can predict who, which cars, are in the best position to move out of the way to calculate the safest route for everyone. Fundamentally, these technologies exist today.
I hteli smo da uradimo - kako najbolje obavestiti svakoga? Kako da ovi automobili šapnu: "Moraš da se skloniš s puta?" To zavisi od dve stvari: pod jedan, mogućnosti automobila i pod dva, mogućnosti vozača. Ako jedan čovek ima stvarno odličan automobil, ali priča na telefon ili već radi nešto, verovatno nije u najboljoj poziciji da reaguje u hitnom slučaju. Počeli smo poseban deo istraživanja gde smo modelirali stanje vozača. Koristeći komplet od tri kamere sada možemo otkriti da li vozač gleda napred, u stranu, dole, da li telefonira ili pije kafu. Možemo predvideti nesreću i možemo predvideti ko i koji automobili su u najboljoj poziciji da se sklone i izračunaju najbezbedniju putanju za sve. U osnovi, ove tehnologije danas postoje.
I think the biggest problem that we face is our own willingness to share our data. I think it's a very disconcerting notion, this idea that our cars will be watching us, talking about us to other cars, that we'll be going down the road in a sea of gossip. But I believe it can be done in a way that protects our privacy, just like right now, when I look at your car from the outside, I don't really know about you. If I look at your license plate number, I don't really know who you are. I believe our cars can talk about us behind our backs.
Mislim da je najveći problem sa kojim se suočavamo naša volja da podelimo svoje podatke. Mislim da je to veoma uznemiravajuća zamisao, da će nas posmatrati naši automobili, o nama pričati sa drugim automobilima, da ćemo putem ići u moru tračeva. Ali verujem da se to može uraditi na način koji štiti našu privatnost, kao sada, kada pogledam vaš automobil spolja, zapravo ne znam ništa o vama. Ako pogledam broj vaših tablica, zaista ne znam ko ste vi. Mislim da naši automobili mogu da pričaju o nama iza naših leđa.
(Laughter)
(Smeh)
And I think it's going to be a great thing. I want you to consider for a moment if you really don't want the distracted teenager behind you to know that you're braking, that you're coming to a dead stop. By sharing our data willingly, we can do what's best for everyone.
Mislim da će to biti sjajna stvar. Želim da za trenutak razmotrite da li zaista želite da rastrojeni tinejdžer iza vas ne zna da kočite, da stajete u mestu. Ako voljno delimo svoje podatke, možemo uraditi ono što je najbolje za sve.
So let your car gossip about you. It's going to make the roads a lot safer.
Pustite vaš automobil da trača o vama. To će puteve učiniti puno bezbednijim.
Thank you.
Hvala vam.
(Applause)
(Aplauz)