Showing posts with label general motors. Show all posts
Showing posts with label general motors. Show all posts

Monday, August 30, 2010

Augmented Reality Will Change How We See The World

Augmented reality is the technology of superimposing virtual images over top of one’s view of the real world. One of the earliest examples is the yellow first down line in televised American football games, which has existed since 1998. The technology typically works by recognizing a certain object in the real world, which triggers a computer to impose a virtual image into the viewer’s line of vision. In the case of the first down line, it operates through a television screen, but any type of screen will do. Augmented reality applications already exist for computers, smartphones, special-design glasses, and airplane windshields.

General Electric has a very cool interactive demonstration of augmented reality with which you can experiment on your computer. In this demonstration, an ordinary computer webcam scans the room. GE’s program detects a certain symbol (on a printed piece of paper), and displays a virtual image over top of it. You can hold the paper from almost any angle. If you move the paper or change the angle, the virtual image will follow your movements so that it constantly appears that the virtual image is on top of the symbol.

But although this is certainly a cool trick, does augmented reality have any useful applications beyond keeping us mildly amused? Any good augmented reality application has at least two components: 1) Recognizing a real-world object to "trigger" the application to do something, and 2) Superimposing a virtual image into the user's line of sight. Imagine you are walking the streets of an unfamiliar city looking for a place to eat lunch. As you walk by restaurants with your augmented reality glasses on, you see the restaurants’ average Yelp rating (and the most recent reviews) hovering just above the buildings. Based on this data, you select the restaurant you want to eat at. This would certainly be more convenient than looking up each restaurant individually as you walked by.

Augmented reality software could be installed in the windshields of automobiles (as it already is in airplanes) to dramatically improve safety, according to General Motors. The technology could scan the car's surroundings for possible hazards. Any object deemed dangerous, such as a deer on the side of the road, could flash on the windshield to draw the driver’s attention to it. In conditions of poor visibility, an augmented reality windshield could help point out the edges of the road, the lines on the road, and any important road signs.

BMW has another example for how augmented reality might work in automobiles in the near future. Suppose that you wanted to repair something in your car, but knew very little about how to go about it. Rather than take it to a mechanic, you could put on your augmented reality glasses, and get step-by-step instructions for how to repair it yourself. The software in your glasses would recognize the parts of your car as you looked at them, and create a detailed illustration showing you exactly what you needed to do in real-time.

The tutorial applications of augmented reality are endless. In addition to repairing a car, augmented reality could help teach people how to perform a wide variety of tasks, from processes as simple as cooking to those as complex as open-heart surgery. Eventually, even augmented reality glasses will become obsolete, as computing power becomes cheap enough to fit inside regular eyeglasses or contact lenses to project virtual images directly onto the user’s eye.

Futurist Ray Kurzweil envisions a future in which contact lenses come with several different viewing modes, depending on what the user wants to see at any given moment. There would be a “normal mode,” which simply displays the real world as we currently observe it. There would be an “augmented mode,” in which the user’s augmented reality applications superimpose virtual images over top of the user’s regular view of the world. Finally, there would be a “blocking mode,” in which the real world was not displayed at all, allowing the user to become fully visually immersed in a website, book, movie, or computer game.

Some of the earliest applications of augmented reality are already being rolled out on the iPhone and Android. This technology seems poised to explode into mainstream use over the next 10-15 years, and will quite literally change how we see the world.

PREDICTIONS:

By 2013 – Useful augmented reality applications exist on PCs and/or tablets to allow shoppers to virtually try on clothing before purchasing it online.

By 2017 – At least one-third of all smartphone and/or tablet users have an augmented reality application to project virtual images over the real world, as seen through their screen.

By 2021 – Augmented reality is routinely used to train people how to perform process-based tasks such as cooking, dentistry, surgery, furniture assembly, factory work, and/or auto repair.

By 2023 – At least half of all new, non-driverless automobiles in the US have augmented reality technology in the windshield for safety and/or navigational purposes.

By 2028 – Augmented reality contact lenses exist which can place virtual overlays of the world directly onto the wearer’s eye, or block out the real world altogether if the wearer desires.

Saturday, May 8, 2010

The Future of Automobiles - Driverless Cars

In 2007, the Defense Advanced Research Projects Agency (DARPA) sponsored an event called the Urban Challenge. Recognizing the utility of self-driving vehicles for military purposes, DARPA invited 11 teams from universities and corporations to build vehicles that would be able to navigate a course through a makeshift city without any human interference. Furthermore, the teams would have to obey all the traffic laws and avoid any collisions with other vehicles, buildings, or obstacles. Six of the teams were able to complete the contest, led by Carnegie Mellon University.

The most successful self-driving cars use cameras and lasers mounted on every side of the car to “see” their surroundings, and send the images to a computer in the car to process them in real time. The technology has already advanced to the point where computers can read road signs, detect other fast-moving objects (like cars), and generally obey the traffic laws. General Motors has announced that they will start testing driverless vehicles in 2015, and hopes to have them on the road by 2018.

The biggest technical obstacle that still needs to be overcome involves dealing with unexpected situations. The driverless car prototypes, such as Carnegie Mellon’s vehicle “Boss,” are fairly good at recognizing and obeying stop signs and traffic lights. However, these vehicles are built with the assumption that all other cars on the road will obey the laws as well. Most humans can hit the brakes if another car runs a red light or if an animal runs out in front of our car, but unfortunately the driverless cars are not quite there yet. However, with the speed at which driverless technology is progressing, it seems very likely that this problem will be overcome in the next few years, and self-driving cars will be able to react at least as well as human drivers.

The impact that driverless cars will have on society will be nothing less than transformative. The biggest revolutions will occur in safety and lifestyle.

In the United States, over 40,000 people are tragically killed each year in car accidents, 95% of which are due to human error. Driverless cars thus offer us the opportunity to save 38,000 lives every year. When the technology matures, the computerized systems in our cars will have reflexes thousands of times quicker than human drivers, and will be able to scan all around the car at all times to identify any potential dangers.

Driverless cars will also improve our lifestyle, by reducing commute times. Driverless cars will be able to identify any road delays via the internet, and plan an alternate route to avoid getting stuck in traffic. Furthermore, when driverless cars become ubiquitous, there will be no need for individuals to own cars. When cars can drive themselves, why would I spend thousands of dollars on a car that will sit unused in a garage or parking lot for most of the day? It would be much more efficient for cities to develop networks of public cars that could drive themselves to pick people up as needed. If I needed to go across town, I could use my phone to order the nearest car to pick me up. If I needed a car to pick me up at the same place and time every day (such as for my morning commute to work), I could schedule this as well. These networks of driverless cars could operate similar to taxi cabs, but much more efficiently, safely, and cheaply.

The biggest long-term obstacles to driverless cars are liability laws. While driverless cars could save up to 38,000 lives per year that would otherwise be lost due to human error, what happens if 1,000 lives are lost due to computer error? Under our present liability system, the auto manufacturers would lose their shirts. Unless these laws are changed, it will be extremely difficult from an economic standpoint for auto companies to roll out self-driving vehicles in the United States for commercial use.

I am optimistic that this problem can be overcome. As soon as the utility of driverless cars becomes clear (probably by 2015-2020), there will likely be a legislative push to limit the liability for auto manufacturers. When that happens, we will enter the era of automated transportation and the world will be forever changed.

PREDICTIONS:
By 2020 - Driverless cars are commercially-available and street-legal somewhere in the United States.
By 2027 - New driverless cars outnumber new cars requiring at least some human control, in the US market.
By 2035 - Driverless cars are widely perceived as safer than human drivers. Somewhere in the United States, it is illegal for humans to drive.