Google self-driving car bears 'some responsibility' in accident for first time ever
Though Google's automated vehicles have been involved in a handful of minor accidents on the road, those were all caused entirely by other, human drivers. But that changed with an accident in California earlier this month.
The incident was revealed in documents filed with the California Department of Motor Vehiclesm unearthed by by Re/code on Monday. They reveal that a self-driving Lexus RX450h hit a municipal bus in Mountain View.
The DMV's accident report states that the self-driving Lexus was attempting to merge back into traffic and avoid some sandbags on the road in front of a storm drain. The Google-controlled vehicle apparently saw the bus in its rear view mirror and assumed it would stop or slow down, but instead, it kept going.
In all, it was a low-speed accident: Google's vehicle was moving at just 2 miles per hour, while the city bus was traveling at 15 miles per hour. There were no injuries, and the report only made mention of damage sustained by Google's car.
After news of the crash surfaced online, Google offered its own account of the accident, saying that similar incidents happen "between human drivers on the road every day."
"This is a classic example of the negotiation that's a normal part of driving -- we're all trying to predict each other's movements," the company said. "In this case, we clearly bear some responsibility, because if our car hadn't moved there wouldn't have been a collision. That said, our test driver believed the bust was going to slow or stop to allow us to merge into the traffic, and that there would be sufficient space to do that."
On public roads, Google's self-driving cars were previously known to have been involved in 17 different accidents. But in each of those incidents, human drivers were said to be at fault, making the Feb. 14 incident the first time Google's vehicle can shoulder some of the blame.
While Google's autonomous vehicle efforts are out in the open, Apple's own "Project Titan" is a secret development said to be underway not far from the company's corporate headquarters. AppleInsider's own sources and research have indicated that the bulk of this development is underway in a series of buildings in the city of Sunnyvale.
Specifically, Project Titan is said to be based out of a building known internally as "SG5." It's there that a company by the name of SixtyEight Research has been operating, prompting speculation that it could be a shell corporation used by Apple to fly under the radar.
Reports have suggested that Apple is hoping to put its own vehicle on the road by 2019, but that the first-generation model won't be a self-driving car. Autonomous capabilities are said to be a more ambitious, longer-term goal for Apple?--?something that could be difficult and potentially dangerous to implement, as evidenced by Google's accident in Mountain View.
The incident was revealed in documents filed with the California Department of Motor Vehiclesm unearthed by by Re/code on Monday. They reveal that a self-driving Lexus RX450h hit a municipal bus in Mountain View.
The DMV's accident report states that the self-driving Lexus was attempting to merge back into traffic and avoid some sandbags on the road in front of a storm drain. The Google-controlled vehicle apparently saw the bus in its rear view mirror and assumed it would stop or slow down, but instead, it kept going.
In all, it was a low-speed accident: Google's vehicle was moving at just 2 miles per hour, while the city bus was traveling at 15 miles per hour. There were no injuries, and the report only made mention of damage sustained by Google's car.
After news of the crash surfaced online, Google offered its own account of the accident, saying that similar incidents happen "between human drivers on the road every day."
"This is a classic example of the negotiation that's a normal part of driving -- we're all trying to predict each other's movements," the company said. "In this case, we clearly bear some responsibility, because if our car hadn't moved there wouldn't have been a collision. That said, our test driver believed the bust was going to slow or stop to allow us to merge into the traffic, and that there would be sufficient space to do that."
On public roads, Google's self-driving cars were previously known to have been involved in 17 different accidents. But in each of those incidents, human drivers were said to be at fault, making the Feb. 14 incident the first time Google's vehicle can shoulder some of the blame.
While Google's autonomous vehicle efforts are out in the open, Apple's own "Project Titan" is a secret development said to be underway not far from the company's corporate headquarters. AppleInsider's own sources and research have indicated that the bulk of this development is underway in a series of buildings in the city of Sunnyvale.
Specifically, Project Titan is said to be based out of a building known internally as "SG5." It's there that a company by the name of SixtyEight Research has been operating, prompting speculation that it could be a shell corporation used by Apple to fly under the radar.
Reports have suggested that Apple is hoping to put its own vehicle on the road by 2019, but that the first-generation model won't be a self-driving car. Autonomous capabilities are said to be a more ambitious, longer-term goal for Apple?--?something that could be difficult and potentially dangerous to implement, as evidenced by Google's accident in Mountain View.
Comments
a private citizen's car runs into a Government vehicle and the Government official says that it is the citizen's fault.
i call bullshit!
i have seen this in real life involving real people. the citizen received a ticket by the officer who caused the accident. the citizen protested and the duty officer arrived and issued the citizen another ticket. the citizen protested and the highway patrol was called. the highway patrol officer gave the officer the ticket and the judge suspended the officer's license for 6 months. quite funny. I guess google will have to program the ability to dispute a ticket with their cars. If corporations are people, then cars are people too!
With a solid record of the incident, there's plenty of evidence for insurance companies (and a court, if necessary) to review and come to a decision. And possibly a better one than would be reached if it was two human drivers and a lot of possibly-inaccurate testimony.
I just bought a new car and it has every piece of high-tech systems for self-parking, collision-detection, and lane-drifting-correction abilities that is almost overwhelming. It drives the car more than I do.
I definitely see a time where automotive manufacturers will implement some kind of common automotive-network system where the cars communicate with each other and can therefore prevent collisions.
if there were no human drivers, these manufacturers could simply get together and come up with standards as to what cars would do in every circumstance. Under these conditions, there would be no accidents that were caused by the cars themselves. The problem right now is that a computer can't predict 100% of the time what a human driver would do. So there are times, like this one, where the car will be in error.
Wait..what?
In this case, if the automated bus had indicated it's decision to the car and vice-versa, there would be no accident at all.
People who think mostly automated driving is far in future will be surprised, it's coming real soon; fully larger scale automated will find it's way on some freeways by 2020.
The obstacle right now are legal and regulatory, not technology.
Cars will not need to communicate with each other to avoid each other. Computer-driven cars are inherently more predictable (by several orders of magnitude) than human-driven cars. If computer-driven cars can avoid causing collisions with human-driven cars (and they have already proven that they can), then avoiding collisions with computer-driven cars is trivially easy. Sensing what the other vehicle is doing and knowing something about its programming obviates any need for communication.jason98 said:
No, computers are much better than humans at avoiding accidents.
Cars will not need to communicate with each other to avoid each other. Computer-driven cars are inherently more predictable (by several orders of magnitude) than human-driven cars. If computer-driven cars can avoid causing collisions with human-driven cars (and they have already proven that they can), then avoiding collisions with computer-driven cars is trivially easy. Sensing what the other vehicle is doing and knowing something about its programming obviates any need for communication.jason98 said:
No, computers are much better than humans at avoiding accidents.