NVIDIA typically has their keynote before CES starts and it is always entertaining. This year the focus was largely on autonomous driving and NVIDIA, who leads in this segment, announced new relationships with VW, Uber, Aurora, Baidu, and others along with advancements to their platform for trucks, driverless cabs, and, of course, self-driving cars. At this point no other technology vendor is as well positioned to ride this wave and the only real risk to the firm’s ability to dominate the segment is the insular nature of the car companies and their tendency to favor their own technology over third parties. NVIDIA is also the only company that seems to get that this technology, while initially targeted vehicles, could be used in drones (including people carrying drones), robotics, and other highly automated projects. It is certainly possible that the future of home automation could eventually be tied to their far more intelligent platforms.
But, what I thought was especially interesting, was their new vision for what these cars will be like as they are one of the first to pivot from the poorly received idea of lost driver control, to Guardian Angel concepts and turning the car into more of a robotic partner than a 4-wheeled elevator. I think this is critical because surveys currently show drivers are anything but excited about this innovative technology and that will need to change if it is to be successful.
I first heard of this concept years ago at a NVIDIA event where Toyota’s CEO was speaking. He was concerned that autonomous cars would take the fun out of driving and collapse the automotive market. This is a concern I share. To address this, he spoke to two concepts Chauffeur, which is the 4-wheeled elevator concept, and Guardian Angel. This last allowed people to seem to drive their own cars but prevented them from making mistakes. The example they showcased at the NVIDIA event was here a stop light turned green but the throttle on the autonomous car was disabled because the car saw that another vehicle was going to run the light and moved to prevent what otherwise would have been a deadly crash. Another example was a woman trying to get out of her car when a bicycle was coming by her car, it locked the door making sure she didn’t injure the bike rider or herself until the bike was no longer a danger and you can picture the same thing if a car was approaching she did now see potentially saving her life.
In effect, the pleasure of driving is maintained but the ability to protect the driver from accidents is also achieved. Kind of like having your cake and being able to eat it too. I can certainly imagine settings in performance cars that would allow the car to help you drive on a track like race car driver with almost no chance of screwing up and wrecking the car.
Car As Companion
This demonstration was the first time I saw the concept of a car as companion. We’ve known for some time that people feel more comfortable if they can talk to the car driving them and if the car keeps them informed about what it is doing when it is in control. But the concept of the car knowing where you are, unlocking doors or the trunk as you approach, through voice command (after authenticating it is you), and being able to recognize you on sight. I can imagine this could evolve into the car knowing where you are and what your situation is so that if you walking toward the exit of a store burdened with packages, if certain conditions were met, it would be waiting at the curb for you and open the trunk so you didn’t have to hike all over the parking lot trying to find it, or meet you at baggage claim just as you come out of the airport (though, I expect, TSA may have issues with lots of driverless cars pulling up to airports). I can even see a future where, should you fall on your way to or from the car, the car would come to your rescue, puck you up, and take you to the hospital. (Something a lot of us could use this year given how icy some of the parking lots and sidewalks are).
This was just cool. Self-Driving cars must be trained and while that training can be passed from one car to another, if there is a glitch, the only way you’d have likely found out was on the road and that is unacceptable. NVIDIA showcased an electronic Self-Driving car simulator that you could plug the brain of the car into and, at hyper speed, have the car virtually drive thousands of miles in minutes validating the car will work as intended.
Being me, I saw this, and it occurred to me that not only would this better assure the performance, particularly for early self-driving cars while training is limited, but it would make a fun strategy game. You see you can modify the environment and the virtual car being driven.
So, the game I envisioned is one where the game AI set up scenarios with weather, varied driving surfaces, and maybe things like attacking Zombies or Aliens and you configured the virtual autonomous car with weapons, engines, wheels, tires, defenses, and selected the personality of the AI driver(s) and then observed the result against something like a capture the flag goal. You could even have virtual car vs. car duals and progress, or in game purchases, would allow you to update your car’s gear.
I think NVIDIA is on to something because a car that takes driving away from you isn’t that attractive, but a car that protects you and your loved ones, and actively moves to make your life easier (and can come to rescue you) would be a very attractive feature set. You can choose to drive or not and, when you drive, you become a flawless driver even if you wonder onto the track (granted I tend to doubt they’ll let you race with this technology but the car could teach you how to drive a particular track fast safely). Rather than taking the pleasure of driving from you, this would effectively turn your car into a companion or even a well-trained pet. I can now imagine walking up to my car saying, “who’s a good boy?” Not that I’d EVER do that…
I can also imagine a future Autonomous car doing this keynote, because, I’ll be, it would rehearse…