Once self driving cars reach a certain penetration level I wonder how long it'll take the remaining human drivers to start taking advantage of that fact by assuming most cars will automatically stop for them when they try stupid maneuvers like this one.<p>I commute on foot in Boston and I regularly witness pedestrians cross streets in front of traffic when they know they can get away with it, for example, when they're part of a crowd of pedestrians or the traffic is moving slowly. Every time I see that I wonder how much worse it'll be when they don't have to worry about being hit. Our sidewalks are narrow and crowded, will pedestrians just move to the streets?<p>Based on our geography, drivers and pedestrians I fully expect my city to pioneer the field of taking advantage of self driving car's automated accident avoidance systems.
The thing that strikes me in that video is that the Tesla is driving really fast considering the right lane is completely stopped.<p>A human driver would never be driving that fast. Taking into account familiarity with typical human behaviour, if you're in a lane that is moving well, you will expect at least some drivers from a slow lane to try to squeeze their way into the faster moving lane. Not to mention watching out for facial expressions on drivers in the opposite moving lane, hoping to take advantage of slow traffic to make left-turns into oncoming traffic. A focused human driver will be taking all these things into consideration when operating the vehicle.<p>I can easily imagine a scenario where a car turns into Tesla's lane and the Tesla not having enough reaction time to avoid collision.<p>Bottom line is, what I'm seeing in this video is a Tesla autopilot not being smart enough to regulate its speed according to conditions on the road and having to resort to some emergency braking maneuver to avoid a potentially catastrophic collision.
This not the effect of Autopilot, Tesla calls it 'Automatic Emergency Braking'.<p>Most modern cars come with similar collision avoidance systems. Volvo introduced it first(?) in 2008.
I'm not convinced that this is an instance of autopilot "saving the day", as the title of the driver's video would suggest.<p>The driver states in his video description that he was "watching stopped traffic to my right". Think how you would react were you in full control of a car in that instance. You have a virtually stopped lane of traffic to your right, you see cars braking about 100 ft ahead of you. Wouldn't you start slowing down? I would, as clearly travelling as fast as the driver is where all surrounding traffic is slowing or stopped doesn't make sense, however the driver doesn't appear to slow down at all, by his own admission focused on the cars to his right. Was he legitimately distracted by cars to his right, or was he depending on the car's situational awareness? I'm leaning towards the latter, what I believe is an affordance offered by autopilot systems.<p>I think we'll start hearing and seeing many such "success" stories of close calls involving autopilot, but I'm not sure how many of these will be incidents of autopilot truly saving an otherwise attentive driver by reacting faster/better than the attentive driver. It seems to me that autopilot is going to be both cause and saviour in far too many cases.
The NHTSA is talking about requiring at least radar-controlled braking on new cars in a few years.[1] It's been working well in high-end cars, and it's not that expensive. It can't prevent all collisions, but it will almost totally prevent rear-ending the car in front in traffic. Most of the major automakers are on board with this.[2]<p>Earlier thinking was only to apply the brakes automatically as a collision-mitigation strategy, so that there would still be a collision, but a less severe one, after which the air bags would probably fire. But the technology got better, and now it looks reasonable to go to automatic braking sooner and prevent many collisions entirely.<p>This will increase freeway capacity. The biggest cause of freeway delay is accidents, most of which are rear-end collisions.<p>[1] <a href="http://www.detroitnews.com/story/business/autos/2015/06/08/ntsb-urges-us-mandate-advanced-braking/28702151/" rel="nofollow">http://www.detroitnews.com/story/business/autos/2015/06/08/n...</a>
[2] <a href="http://money.cnn.com/2015/09/11/autos/automatic-braking-nhtsa-iihs/" rel="nofollow">http://money.cnn.com/2015/09/11/autos/automatic-braking-nhts...</a>
Here is what I want.<p>I want the car to avoid the collision, record the whole incident, and submit it as evidence to penalize the other driver.<p>The penalties could take the form of a lawsuit. Or could take the form of tagging the incident, and submitting the video where the other person's insurance company can automatically pick it up, examine the evidence, and then change rates.<p>That way we not only prevent accidents, but we start to address the minority of reckless drivers that put all of us at risk.
This is where autopilot (and self-driving car technology in general) really shines: faster-than-human reaction time. The driver didn't have enough time to even honk, let alone stop the car.
My car was totaled in a near identical situation last month. And this was avoided with only automatic emergency braking... I long for the day when fully self-driving cars are the norm on roadways.
What kind of Uber driver owns a Tesla? Is the guy just doing it for fun/to meet people or is it actually feasible to make payments on a Tesla with an Uber driver's salary?
Neat. Question: what happens if there's a car in back of the Tesla? Does does the Tesla disregard him and whether or not he has time to stop as well?<p>Edit: Certainly the top priority is avoiding the car in front of you, and you should slam on the brakes to do it. However, once you stop, the Tesla should release the brakes (if there's an imminent collision from the rear) to minimize the impact to the Tesla driver and the car slamming into the back of him. Would be cool if the Tesla engineers added this (if feasible).
It mentions in the video that the weather was dry. Is weather a variable that the autopilot system takes in to account? For example, if it was raining would the car have slowed down to a speed where it knows it can avoid obstacles taking in to account the extra stoppage distance required?
I wonder if all of this tesla-autopilot hype might increase their chances of being in a collision? At least initially.<p>I (regrettably) found myself leering over into a tesla driver's console the other day as he drove past to see if he was using autopilot or not and started to think of how dangerous it probably is to have everybody around you looking at you and not the road.<p>I guess at the very least tesla's autopilot will have some good practice early on.
I know I would definitely have been unable to prevent the accident. I didn't see the car until it was on the road.<p>Reminds me of the couple of close ones I had either in my car on motorcycle. On the moment you don't think, you just do but after it's a very chilly feeling of realizing that your life was just being played over a couple of seconds.
I know it's not in HN nature, but I absolutely loved the fact one of the top comments on youtube is: "FU<i></i> YEA TECHNOLOGY!". It actually means that people outside tech environment appreciate course of events that is happening in the tech industry.
:sarcasm: Nothing to see here just some sensors, actuators and few lines of embedded code in action. Thousands of kids do it with Lego mindstorm every days.
Indian roads, especially in towns, are where all self-driving cars will fail.)<p>Push the horn constantly and go to opposite lane is very common strategy, for example.
Url changed from <a href="http://fortune.com/2015/10/29/tesla-autopilot-uber-crash/" rel="nofollow">http://fortune.com/2015/10/29/tesla-autopilot-uber-crash/</a>, which points to this.