Tesla driver dies in first fatal crash while using autopilot mode | Hacker News Reader

Tesla driver dies in first fatal crash while using autopilot mode

theguardian.com

57 pointssummerdown212 comments

Open articleView on HN

Looks like the truck had a truncated view of the car because of a grade.

http://imgur.com/a/lFRgf

Elon says the radar ignores what it thinks are overhead signs.

https://twitter.com/elonmusk/status/748625979271045121

More discussion:

https://m.reddit.com/r/teslamotors/comments/4qnu4a/a_tragic_...

"The top of Joshua Brown’s 2015 Tesla Model S vehicle was torn off by the force of the collision. The truck driver, Frank Baressi, 62, Tampa was not injured in the crash. The FHP said the tractor-trailer was traveling west on US 27A in the left turn lane toward 140th Court. Brown’s car was headed east in the outside lane of U.S. 27A. When the truck made a left turn onto NE 140th Court in front of the car, the car’s roof struck the underside of the trailer as it passed under the trailer. The car continued to travel east on U.S. 27A until it left the roadway on the south shoulder and struck a fence. The car smashed through two fences and struck a power pole. The car rotated counter-clockwise while sliding to its final resting place about 100 feet south of the highway. Brown died at the scene. Charges are pending."

source: https://www.levyjournalonline.com/police-beat.html

given the above, I suspect that even if the Model S' brakes kicked in, they still would not have saved him. Unless the driver was asleep at the wheel or otherwise distracted.

View of intersection: https://goo.gl/maps/nmbNuGMkbQs

> The first paragraph notes that this was Tesla’s first known autopilot death in some 130 million miles driven by its customers. “Among all vehicles in the US, there is a fatality every 94 million miles,” the company then notes.

This statistic is a bit disingenuous, since the 1 fatality every 94 million miles is across all road types, where as Tesla's autopilot is only used for highway driving, mile for mile the safest way to drive.

Very interesting that you suggest highway driving is the safest part (I assume in the US). In Australia it is country driving that is the most dangerous. An area the Telsa self driving feature would be very suited.

http://www.mac.sa.gov.au/campaigns/country-driving

This is why i'm skeptical of tesla rushing out autonomous features. i think augmented driving being confused by customers as fully autonomous is extremely dangerous. teslas rush to market and resulting accidents could set back the whole industry.

I also don't understand how tesla can make the leap to fully autonomous vehicles learning/using training data from the semi-autonomous vehicles that are in the customers hands? google has driven 1.5 million miles fully autonomously (with some human oversight) on city streets how many has tesla?

another thought: if tesla does get to fully autonomous vehicles... wouldn't that destroy their sales? its generally agreed upon the demand for car ownership decreases dramatically with self driving vehicles. maybe theyre not afraid to cannibalize themselves akin to apple?

Even if the software was blameless, I could see an incident like this setting back acceptance of autonomous driving a couple of years.

Brad Templeton argues a perpendicular truck would look stationary to radar and be filtered out, so use LIDAR.

http://ideas.4brad.com/man-dies-while-driven-tesla-autopilot

The reason they couldn't use LIDAR on the Tesla is because Tesla is literally the coolest car on the road, and the LIDAR protrusion on the top of the car would completely reverse that. It would go from the coolest car to the dorkiest car.

There isn't a way to just stick the LIDAR somewhere unobtrusive. So this is the thing that, strangely, almost no one (aside from myself) seems to have noticed: LIDAR is fundamentally incompatible with the current Tesla brand. And that was especially the case when there was only really the super high-end 'cool' model of Tesla vehicle. But still, the way the vehicles are marketed and designed today, they cannot add a functional LIDAR without that fundamental change.

This is a little bit ridiculous, but the reality is that at some level, most average adult consumers out there are not really much more sophisticated than the average kid in your 7th grade English class. So it just comes down to the fact that the thing on the top doesn't look cool.

Now, Elon Musk (and most Tesla fans) will never admit that.

You could add it as a fin, like BMW antennas. LIDAR is expensive and mechanically unreliable: you can't steer light electrically, so you need either a spinning motor or a phased array (under research). And isn't really needed after optical, radar and sonar.

I don't understand how it would look stationary. The car is moving towards the truck, therefore it would not appear stationary on radar, right?

Known car velocity is subtracted out. The bigger issue is why it wouldn't halt if a stationary object is in your lane. My guess is the sensors had too narrow a vertical FOV to detect a high tractor trailer.

Reply on news.ycombinator.com