Blue Line

News
Autonomous cars are approaching

January 29, 2016  By Tom Rataj


1063 words – MR rataj mar.jpg, rataj2 mar.jpg

Autonomous cars are approaching

by Tom Rataj

No longer entirely the stuff of science fiction, autonomous cars are slowly making their way into the mainstream. Many modern cars already have some autonomous features such as parking assistance, collision avoidance and lane-departure control.

Advertisement

The few fully autonomous research vehicles are mostly confined to closed driving circuits. The exception is Google’s fleet of 23 self-driving cars, which have already logged over 1.2 million kilometers on public roads.

Tesla, the world’s largest all-electric car manufacturer, recently added more autonomous driving capabilities to its Model-S line. The cars can now travel in traffic, at highway speeds, without steering or other inputs from their human “driver.”

The truck division of Mercedes-Benz has also successfully demonstrated an almost 100 per cent autonomous transport truck that requires virtually no driver input.

In his most recent state-of-the-union address, US President Obama pledged $4 billion over 10 years to accelerate the development and implementation of autonomous vehicles and their technologies. Part of the motivation is to reduce the high number of road fatalities by eliminating the most common cause: human error.

Widespread availability of autonomous cars may be here in as little as four years if one is to believe people like Elon Musk of Tesla and SpaceX, although the vehicle price-point will be far above affordable levels for most people during the first few years.

It will eventually be as simple as getting into your car and giving a few plain-language directions to the car, such as “Take me to Canadian Tire.”

There will, of course, need to be a huge rethink and adjustment to all sorts of rules, regulations, laws and other legal processes to safely implement this. The effects on policing will also be profound in terms of traffic related management and enforcement.

{How it works}

Autonomous vehicles use a wide range of electronic and other sensing technologies and systems to “see” and recognise their surroundings in much the same way a human driver does. A multitude of systems work together to safely guide the vehicle along its route.

A major part of the technology is “computer-vision” systems, which rely on multiple sensors working together. These may include optical cameras, radar, LIDAR (laser) and infrared. The systems can recognise objects such a people, animals and other vehicles and read road and speed-limit signs.

Additional technologies typically include GPS and odometry, high-quality digital road maps and a live cellular connection to road and traffic information and conditions, local speed limits and a wide variety of other information to assist the car in its decision making processes.

Some systems use vehicle-to-vehicle (V2V) communication technologies that establish automatic awareness. Each vehicle is aware of all the others in its vicinity, sending and receiving information about behaviour, speed, direction and other factors.

Vehicle to Infrastructure (V2I) technologies – systems which use electronic or computer-vision friendly markers along roadways for navigational purposes – are also used.

{History}

Work on autonomous vehicles has actually been ongoing since the early days of the automobile, although much of the early work would better be classified as “automated” as opposed to autonomous.

These were essentially just dumb systems with very rudimentary decision making capabilities. Unmanned robotic transport systems used in many large factories and warehouses are a good example.

Work on fully autonomous vehicles only really became a reality in the 1980s when computer technology and sensors progressed enough to allow vehicle control systems to ‘think’ and react to multiple problems in an uncontrolled environment.

In addition to university research projects, every major car manufacturer and many vehicle technology and equipment suppliers are all feverishly working on developing vehicles, systems and equipment for autonomous vehicles. The QNX division of BlackBerry is a world leader in this field.

{Legislation}

Several provinces and states have already passed or are developing legislation to deal with autonomous vehicles.

Both the Society of Automotive Engineers (SAE) International and The National Highway Traffic Safety Administration (NHTSA) have proposed similar five level autonomous vehicle classification systems. They being with a vehicle controlled entirely by a person and end with a vehicle completely controlled by autonomous systems.

Several European cities have begun allowing testing of completely autonomous speciality vehicles on public roads.

{Legal implications}

I can almost hear this explanation: “But officer, I wasn’t driving, my car was on auto-mode, and I don’t know how the crash happened!”

Besides being immediately suspicious, how will police and courts deal with this? Perhaps some kind of electronic recording and auditing technology and new laws will be needed.

All sorts of smarts and artificial intelligence protocols will be required to deal with extraordinary situations. A version of Isaac Asimov’s Three Laws of Robotics may be useful to ensure autonomous vehicle protocols value human life above all else and make decisions accordingly.

How will a vehicle deal with crashes that are beyond its control or decide, deciding in a split second what to do and which ethical or moral rules apply?

{Advantages/disadvantages}

There are numerous potential advantages to the widespread use of autonomous vehicles. Chief among them will be the reduction of traffic fatalities and injuries. US estimates suggest that extensive use of autonomous vehicles could eliminate 25,000 road fatalities each year.

Other advantages will include reduced congestion (because vehicles can safely drive closer together), less need for traffic enforcement, removal of current constraints such as age and driver sobriety and reduced vehicle theft.

There are also numerous disadvantages and problems that will need to be worked out, including the need for multiple system redundancies in case of a computer crash and vehicles potentially being hacked or otherwise remotely commandeered.

Interestingly, as of mid-2015 Google’s self-driving car fleet had been involved in only 14 minor collisions. All but one were caused by people driving conventional cars, with the one at-fault collision occurring when a person was manually driving a Google car.

I believe people in their mid-50’s like me will be the first generation that won’t need to give up their driver’s licence (and independence) in their 70s and 80s thanks to common and affordable fully autonomous vehicles.

The widespread adoption of fully autonomous cars will be a truly disruptive innovation with widespread impacts on society on many levels. Will vehicle owners still require driver’s licences and need to stay sober while on the road?

“Home, James” will have a whole new meaning!


Print this page

Advertisement

Stories continue below