[NetfreakBombay]

Originally Posted by speedmiester

Autonomous cars is not a new idea, many auto manufacturers have ventured into this idea many years ago. So this has nothing to do with one particular company.

Did any other company come up with a version good enough to be tested on public roads?

Google cars have completed 300 K miles without a single accident. When this technology goes main-steam this would be good for safety.

Even if car is driven by a person, technology can detect and report things like:

1. Driving over speed
2. Drunk Driving
3. Lane cutting
4. Aggressive driving

That should make things safer.

[joslicx]

Originally Posted by NetfreakBombay

Did any other company come up with a version good enough to be tested on public roads?

Almost all of them have! Go thru the Autonomous car wiki page for more details.

Some interesting excerpts:

In 2011, Mercedes-Benz announced its 2013 S-Class will feature an autonomous driving system. At speeds of up to 25 mph, the S-Class will drive itself through slow-moving traffic jams using a series of cameras and radar-based monitoring systems controlling the speed and distance to the vehicle in front

In 2010 VisLab ran VIAC, the VisLab Intercontinental Autonomous Challenge, a 13,000 km test run of autonomous vehicles. Four driverless electric vans successfully ended the drive from Italy to China, arriving at the Shanghai Expo on 28 October 2010. It was the first intercontinental trip ever with autonomous vehicles

In 1995, Dickmanns´ re-engineered autonomous S-Class Mercedes-Benz took a 1600 km trip from Munich in Bavaria to Copenhagen in Denmark and back, using saccadic computer vision and transputers to react in real time. The robot achieved speeds exceeding 175 km/h on the German Autobahn, with a mean time between human interventions of 9 km, or 95% autonomous driving. Again it drove in traffic, executing manoeuvres to pass other cars. Despite being a research system without emphasis on long distance reliability, it drove up to 158 km without human intervention

Google cars have completed 300 K miles without a single accident. When this technology goes main-steam this would be good for safety.

Thats not true. There was an accident involving Google autonomous car though they attributed it to human error!!

[NetfreakBombay]

Originally Posted by joslicx

Thats not true. There was an accident involving Google autonomous car though they attributed it to human error!!

Google car was stationary

It was rear-ended by a Human-driven car.

[joslicx]

Originally Posted by NetfreakBombay

Google car was stationary

It was rear-ended by a Human-driven car.

Well you are talking about a different accident then. See this:

Human error blamed after Google

[rajshenoy]

On a lighter note, We in india are already beyond this.

Atleast here in Bangalore, when i look at some of the morons behind the wheel i definitely feel that Humans cant be driving like this !!!

[NetfreakBombay]

Originally Posted by joslicx

Well you are talking about a different accident then

Even in this case, from the article :

Internet giant says its modified Prius was being manually driven

[joslicx]

Originally Posted by NetfreakBombay

Even in this case, from the article :

And how do we know if Google is not lying? Its not that they are incapable of isint it.

The only fact is, one of the Google cars has been in an accident. Who caused it is unknown. Even Google has not released any report on the same (and they have so many cameras and other instruments inside, surely they would have enough evidence).

[NetfreakBombay]

Originally Posted by joslicx

Who caused it is unknown. Even Google has not released any report on the same (and they have so many cameras and other instruments inside, surely they would have enough evidence).

Drivers of other cars would have field insurance for sure.

Given that this happened in US, insurance lawyers would have easily accessed these if they had any doubt.

[Guna]

Originally Posted by Daewood

What if all the vehicles are on auto-pilot and manually driven vehicles are banned in city roads, like how bullock carts and horse driven carriages are illegal now.

If all the cars are auto pilot, it is a different thing. In that case they don't have to be totally driven by themselves but rather controlled/aided by a centralized system. Concept of interconnected cars is being tried out by another tech company.

[joslicx]

Originally Posted by Guna

If all the cars are auto pilot, it is a different thing. In that case they don't have to be totally driven by themselves but rather controlled/aided by a centralized system. Concept of interconnected cars is being tried out by another tech company.

Wouldnt it make it so easy for any terrorists? Or rogue robots/computer systems? Or am I being too much influenced by Die Hard 4.0/I robot etc.

[mayuresh]

Originally Posted by mayuresh

For road safety, too, let us first see assistive technology than the one that
replaces the driver. Something that helps avoid accidents by alerting the
driver or in extreme cases, taking control and driving to safety automatically.

The point I tried to make in above post was: safety should be the main goal of
technology - driverless or not is only incidental.

Coincidentally I got to read the following article. This is an excellent
article to read for anybody who took interest in this thread. Unfortunately I
could not find a publicly available copy of it. If possible please do check if
any library you have access to can provide a copy to you.

Miad Faezipour, Mehrdad Nourani, Adnan Saeed, and Sateesh Addepalli. 2012.
Progress and challenges in intelligent vehicle area networks. Commun. ACM 55, 2
(February 2012), 90-100. DOI=10.1145/2076450.2076470
Progress and challenges in intelligent vehicle area networks

Some highlights (this is really short and hope I am not violating any norms):

About VAN: Vehicle Area Networks (VANs) will be a significant part of
Intelligent Transportation Systems (ITS) in future. It is expected that VANs
will be ubiquitous by 2016.

Article covers, the goals of VANs and a survey of state of the art which should
be of more interest for this thread. The article also highlights the technical
challenges in achieving these goals which may also be of interest for a subset
of readers.

The ultimate goal of VANs is stated as "to provide an accident-free environment
and move forward the implementation of the zero-accident car"

The survey mentions currently available driver assistance systems and the car
makers who made them. The systems include: collision warnings and mitigation
brakes, lane keeping support, blind spot detection among a few others.

Key components of VAN are divided into following categories:

- In-Vehicle (various monitors and medical assistance equipments etc), further
subdivided into passive (reactive) and active (preventive) safety measures

- V2V (vehicle to vehicle) (data exchange about road condition, safe space,
lane keeping etc.)

- V2I (vehicle to road infrastructure) (auto detection of
signals/obstacles/road conditions and warnings related to them)

- V2B (vehicle to broadband cloud) (driver behavior analysis, fatigue
detection etc., other vehicle data such as speed, location).

[mayuresh]

More from Google on this:

http://www.computerworld.com/s/artic..._s_Sergey_Brin


A video article regarding a mobile app for safety assistance:

http://www.newscientist.com/article/...r-smarter.html

Such effort, or the one like VLANs (see my immediate previous post on this
thread) is what I personally think to be a way forward for improved safety
_before_ more ambitious works like self driven cars.

[apachelongbow]

It's ridiculous that we want technology to replace everything we do. We are living thinking animals and we need to utilise our capabilities. If machines and robots replace everything we do, then should we just breathe eat and exist?


Autonomous cars are a long way from reality, it's one thing to drive around in controlled environment and other thing to expect a robot to drive in the real world day after day and safely. There is a reason why ships and airplanes are not fully automated, robots will never be able to think like a human.

[anujatwork]

Originally Posted by speedmiester

But why haven't we replaced the human pilots. Will you fly an Airline which has no human pilots?

We have indeed completely replaced airline pilots that too on commercial airliners. I have flown domestic through zero visibility wherein both take off and landing was automatic. Pilot just sat watching his wind screen.

Originally Posted by speedmiester

In a real world scenario how efficient will be a computer in driving a car in a city traffic. In a country like India, where there are more potholes than road, people crossing in the middle of the road, cows lying on the road, two wheelers cutting you across and many other situations, can a computer be made smart enough to accommodate for all these variables. Will a person really trust a computer driving his car?

A computer will not drive it basis a program alone. It will collect data from various sensors all over the car and take decision to accelerate/break/turn basis those stimuli. I do agree that a computer will keep a larger margin to avoid error thus your drive will be much slower. For eg. rear parking sensors beep much sooner while you may still have enough distance to back-up if you are an expert driver.

When Metro was launched in Delhi, since that day it is a driver-less train. People psychology is to be scared to ride driver-less so they planted one there just to oversee the auto-pilot or, perhaps, intervene in emergencies. 20 yrs back, all elevators had an operator. Now we step into one at mid night all alone assured that there is a 99.9% chance of not needing help mid-way. It's all a matter of getting used to it.

[speedmiester]

Originally Posted by anujatwork

We have indeed completely replaced airline pilots that too on commercial airliners. I have flown domestic through zero visibility wherein both take off and landing was automatic. Pilot just sat watching his wind screen.

Pilots have not been replaced by computers, they are there to assist in difficult situations as well as to fly at cruising altitude to give the pilots a break. Each flight has a specific flight path that is approved and use various navigational aids to guide them through the flight path. Planes do not have to deal with heavy traffic like we do on roads.
Most of the time during emergencies the first thing the pilot does is to take over the control from Autopilot. Even while approaching a busy airport, the pilots takes over to control the altitude and distance between other planes waiting to land.

When Metro was launched in Delhi, since that day it is a driver-less train. People psychology is to be scared to ride driver-less so they planted one there just to oversee the auto-pilot or, perhaps, intervene in emergencies. 20 yrs back, all elevators had an operator. Now we step into one at mid night all alone assured that there is a 99.9% chance of not needing help mid-way. It's all a matter of getting used to it.

Please don't compare elevators and trains to driving on roads.
The elevator only has set number of variables to look into and it only goes up or down.
Similarly trains can be programmed easily as they run on tracks which don't have to go over potholes, road humps and other things which we encounter regularly on roads. They don't have to deal with other trains on the track in bumper to bumper situation.

BTW you are contradicting yourself in your post, no where have you said that machines or automation have completely taken over transportation and relieved humans from such jobs. There is always some sort of human presence to take over in case of emergencies. Why?, because machines are only as good as their programming and you cannot program every single emergency situation.

I would like to take a different approach to this.
Imagine a bottle consisting of 100 identical tablets which consists of 95 antacid tablets, 4 are medication which can cause severe side effects and 1 poison which can be fatal. How many of us are willing to take a chance of swallowing a single tablet from the bottle.