Even the Robots are Out to Get Us (Uber Self Driving Cars)
“We can confirm that, as the Coalition’s note states, our vehicle operators ‘have been instructed to disengage from self-driving mode when approaching right turns on a street with a bike lane and that engineers are continuing to work on the problem,’†a spokeswoman from Uber wrote in an email.
It’s pretty clear that Uber’s autonomous cars aren’t ready for prime time. Still a work in progress along the lines of Tessa’s auto pilot.
Scary stuff for sure. It is interesting discussion in AI as it is hard to think about everything that can come up, but this just seems like one of those things that is just unfair in life. I don’t want to get into a long debate at this time, but I think there is a prioritization that needs to happen with cars, and they need to be last in line. We as a society should some how rewards those who walk / bike, besides just the value it provides to their own health.
This problem has been solved already by Tesla. They have a YouTube video where the radar detects pedestrians and can easily identify cyclists in motion for sure. It’s just a matter of optimizing the software behavior for it.
I’ll bet Google’s driving car already has solved this.
Uber will figure it out soon but it may take a bit of time. I don’t think it’s scary though I’ll bet there is a clear driver alert to take over. Probably makes it safer actually.
This problem has been solved already by Tesla. They have a YouTube video where the radar detects pedestrians and can easily identify cyclists in motion for sure. It’s just a matter of optimizing the software behavior for it.
I’ll bet Google’s driving car already has solved this.
Uber will figure it out soon but it may take a bit of time. I don’t think it’s scary though I’ll bet there is a clear driver alert to take over. Probably makes it safer actually.
No doubt in the future we will have this figure out. It is the transition period that scares me. All new ‘things’ need their kinks worked out.
I don’t think you need to be so scared of upcoming robot cars.
The number of human-injuring car incidents is miniscule if factoring in the number of miles they have driven on highways, in comparison to human driver error.
I would vastly prefer my chances right now as a cyclist faced with the imminent theat of an autonomous Uber experimental car about to right hook me, than a typical rush hour driver about to right hook me. Don’t even get me started on the texting drivers…
I don’t think you need to be so scared of upcoming robot cars.
The number of human-injuring car incidents is miniscule if factoring in the number of miles they have driven on highways, in comparison to human driver error.
I would vastly prefer my chances right now as a cyclist faced with the imminent theat of an autonomous Uber experimental car about to right hook me, than a typical rush hour driver about to right hook me. Don’t even get me started on the texting drivers…
Scare might be the wrong word or just too much hyperbole. Fwiw, I commute nearly every where by bike so riding in the city is not the problem. I guess, it would just be unfortunate to be a cyclist and be taken out by an autonomous vehicle in that manner. No doubt humans make plenty of errors driving, it is the thought that the humans just left something out in the code that bothers me. The rush to innovate and missing out on important line or sub-program that recognizes cycling behavior.
Saw this on another forum this morning. My expectation is that the end-game of driverless cars will be to force ALL non-automated traffic off the public roadways, which will include bikes.
I don’t think you need to be so scared of upcoming robot cars.
The number of human-injuring car incidents is miniscule if factoring in the number of miles they have driven on highways, in comparison to human driver error.
I would vastly prefer my chances right now as a cyclist faced with the imminent theat of an autonomous Uber experimental car about to right hook me, than a typical rush hour driver about to right hook me. Don’t even get me started on the texting drivers…
Scare might be the wrong word or just too much hyperbole. Fwiw, I commute nearly every where by bike so riding in the city is not the problem. I guess, it would just be unfortunate to be a cyclist and be taken out by an autonomous vehicle in that manner. No doubt humans make plenty of errors driving, it is the thought that the humans just left something out in the code that bothers me. The rush to innovate and missing out on important line or sub-program that recognizes cycling behavior.
The great thing about autonomous vehicles is that if you fix a mistake on one, you fix it on all of them.
Yes it would be unfortunate and sad if a cyclist was hit because the system didn’t recognize him properly. But that would spark even more effort than you’re seeing already - to make sure it never happens again.
Contrast that with human drivers who are all different and way less predictable. As a cyclist, I’ll be super excited when these cars are the norm.
I do think these companies should be sharing the data they collect on car and bike behavior. They can still compete in this arena while working together on safety. I hope they make that happen.
Yes it would be unfortunate and sad if a cyclist was hit because the system didn’t recognize him properly. But that would spark even more effort than you’re seeing already - to make sure it never happens again.
So … would you be willing to take one for the team?
Yes it would be unfortunate and sad if a cyclist was hit because the system didn’t recognize him properly. But that would spark even more effort than you’re seeing already - to make sure it never happens again.
So … would you be willing to take one for the team?
If I have to get hit by a car, I’d rather it be one of those than some idiot texting.
This problem has been solved already by Tesla. They have a YouTube video where the radar detects pedestrians and can easily identify cyclists in motion for sure. It’s just a matter of optimizing the software behavior for it.
I’ll bet Google’s driving car already has solved this.
Uber will figure it out soon but it may take a bit of time. I don’t think it’s scary though I’ll bet there is a clear driver alert to take over. Probably makes it safer actually.
No doubt in the future we will have this figure out. It is the transition period that scares me. All new ‘things’ need their kinks worked out.
Companies are taking very different approaches.
Google has legions of their little Waymo cars doing endless laps of Mountain View (and now Palo Alto). The cars are super-cautious, drive annoyingly slow, and brake at the most minor provocation. Google is interested in perfecting before expanding.
Apple is doing everything on a closed track…because at Apple everything must be secret.
Tesla is putting a 1/2-way technology in the car an leaving the discretion to the driver.
Uber is disregarding regulations and just putting their stuff on the streets in a fairly reckless manner (there’s already a video of an Uber self-driving Volvo running a red light and almost taking out a pedestrian).
This problem has been solved already by Tesla. They have a YouTube video where the radar detects pedestrians and can easily identify cyclists in motion for sure. It’s just a matter of optimizing the software behavior for it.
I’ll bet Google’s driving car already has solved this.
Uber will figure it out soon but it may take a bit of time. I don’t think it’s scary though I’ll bet there is a clear driver alert to take over. Probably makes it safer actually.
No doubt in the future we will have this figure out. It is the transition period that scares me. All new ‘things’ need their kinks worked out.
Companies are taking very different approaches.
Google has legions of their little Waymo cars doing endless laps of Mountain View (and now Palo Alto). The cars are super-cautious, drive annoyingly slow, and brake at the most minor provocation. Google is interested in perfecting before expanding.
Apple is doing everything on a closed track…because at Apple everything must be secret.
Tesla is putting a 1/2-way technology in the car an leaving the discretion to the driver.
Uber is disregarding regulations and just putting their stuff on the streets in a fairly reckless manner (there’s already a video of an Uber self-driving Volvo running a red light and almost taking out a pedestrian).
Sure doesn’t help that Uber lost some giant amount of cash in their latest earnings. I and am sure they are struggling to find a way to eliminate their number one cost, the driver, and make a jump towards profitability. The best way to ensure the bikers are safe IMO is to put Larry and Sergey on the track on bikes. I am sure the engineers will make sure not to hit them.
Apple gave up. Car manufacturing was way harder than they expected, and they weren’t willing to put down the capital.
Most of the majors (Ford, GM, Daimler, etc.) are putting in a lot into semi-autonomous driving. I don’t expect much in city driving within the next decade, but expect that you’ll be able to buy a stock C-Class which rolls along the highway without assistance and knows how to stop for a stray sheep. And avoid cyclists.
Apple gave up. Car manufacturing was way harder than they expected, and they weren’t willing to put down the capital.
Most of the majors (Ford, GM, Daimler, etc.) are putting in a lot into semi-autonomous driving. I don’t expect much in city driving within the next decade, but expect that you’ll be able to buy a stock C-Class which rolls along the highway without assistance and knows how to stop for a stray sheep. And avoid cyclists.
Yeah, I’m guessing you’ll start to see self driving long haul trucks that move between depots via highways. Human drivers will then take over for the “last-mile” from the depots to the destinations.
Figuring out the highway issues are much easier than neighborhoods, and the value of having a six-figure truck not sitting idle 14hrs. per day is high.
This problem has been solved already by Tesla. They have a YouTube video where the radar detects pedestrians and can easily identify cyclists in motion for sure. It’s just a matter of optimizing the software behavior for it.
I’ll bet Google’s driving car already has solved this.
Uber will figure it out soon but it may take a bit of time. I don’t think it’s scary though I’ll bet there is a clear driver alert to take over. Probably makes it safer actually.
mmWave line of sight radars on every corner of the car (like a ton of radar antennae). Today you can put 16 antenna on a single chip in 28 nm CMOS process technology that will do bean forming/beam steering and also have built in dopler so that you can differentiate between static objects, bikes and pedestrians that have different dopler signatures. Embed some Kalman filtering and you can steer the beam where the pedestrian or the cyclist will be “next”. We’ve done all the math 20-40 years go in airborne radar and now we can put them all over cars with all the shrinking of CMOS tecnology and mmWave unlicensed spectrum being available. The big challenge will be regulatory, not technology. the math is there, the semiconductor technology is there…it’s just a matter of bringing it all together (I am probably giving away a half decent startup idea on tri forum, but I digress…)
Apple gave up. Car manufacturing was way harder than they expected, and they weren’t willing to put down the capital.
Most of the majors (Ford, GM, Daimler, etc.) are putting in a lot into semi-autonomous driving. I don’t expect much in city driving within the next decade, but expect that you’ll be able to buy a stock C-Class which rolls along the highway without assistance and knows how to stop for a stray sheep. And avoid cyclists.
Yeah, I’m guessing you’ll start to see self driving long haul trucks that move between depots via highways. Human drivers will then take over for the “last-mile” from the depots to the destinations.
Figuring out the highway issues are much easier than neighborhoods, and the value of having a six-figure truck not sitting idle 14hrs. per day is high.
Long haul trucking should dissappear totally. Like you said, it should all evolve very similar to aviation with larger vessels fullfilling the long distance deliver (that are automated) and regional vessels (which are human driven) doing the local aggregation/disaggregation. I would see this would evolve to more container like traffic which can be put on the larger long haul truck and then sorted locally for human driver enabled localized distribution. Almost like what happens now between trains and trucks but on a different scale. As cyclists, this should not affect us since we don’t ride on interstates.
You guys saw this Budweiser beer gig between Fort Collins and Colorado Springs that we done by a self driving truck?