You May Never Drive a Google Car, but that’s the Point
At this point we’re pretty accustomed to the seemingly regular updates on Google’s progress in creating a safe and working self-driving car. You read about new mapping technology, lasers, car-to-car communication, and an ever-expanding AI. One thing you don’t hear a lot about is the cost of the Google Car. As with anything, the prototype will be more expensive than production models (most likely, but not guaranteed), but how much could the Google Car cost that first group of intrepid early adopters and will they get their money’s worth?
Google Car Technology is Pricey
Research and development of any tech device can be expensive. The same goes for the development of most high-end cars. It’s also a lengthy process to test and make changes to something as complex as a self-driving car. This has been pretty well documented in the development process of the Google car which, back in 2012, was confirmed to have over $150,000 worth of additional equipment attached to a modified Toyota Prius. The latest reports of the program state that Google is now using a modified Lexus R450h. The make and model aren’t as important as the fact that they’re using different types of vehicles—which suggests that they intend to produce software that can be included in various vehicles as opposed to Google attempting to create their own car that utilizes this technology. Changes in models will mean additional programming needs—at added cost.
There is some standard equipment that will be necessary for any self-driving car; actual parts that will need to be attached to self-driving vehicle models. One example is the LIDAR system that the Google car uses to “see” the road, any obstacles, and other vehicles. This hardware is essentially the car’s eyes. A report circulating in 2012 stated that a German manufacturer would supply “a company” (presumably Google) with onboard LIDAR equipment for $250 per vehicle. A more current report states that the car is using a very advanced LIDAR system costing an estimated $70,000. It’s possible that the system is for research purposes only and that a less expensive model might be used once the self-driving car technology is pushed into production, but it will undoubtedly be an additional cost and one that may prove to be a barrier for entry to many drivers (or rather, would-be non-drivers).
Driver and Car Safety
Car safety is obviously important when it comes to automated cars. Knowing when to stop and when to go is one thing, but the car needs to be able to respond to external variables such as reckless drivers. The catch with self-driving cars as they currently exist, particularly the Google car, is that they rely on programming. They aren’t learning computers, they are programmed machines.
Google’s cars are limited to just a few cities because the vehicle has to be programmed with all of the nuances of traffic within a certain area—essentially building out their map like an explorer would. It’s taken quite a bit of time to complete this process and a considerable amount of driving. As a matter of fact, Google recently announced that the fleet had completed a collected 700,000 miles of driving without incident. That’s kind of a big deal.
One of the initial concerns of the program was how the self-driving car would handle so many variables when it came to driving on real-world city streets. The programmers knew early on that programming freeway driving would be simpler than city driving because there were fewer variables. In the city, the car would need to be able to track not only traffic variables, but bikes, pedestrians, animals, and more. That’s a lot of unknowns and a lot of risk. Car safety is a priority, but not nearly as important as public safety.
To compensate for these risks while remaining functional enough to learn to drive in such an environment, programmers set the automated driving mode to be extra cautious. The car will analyze scenarios and then determine whether or not it can proceed (depending on how it understands its current situation and whether or not it can resolve a solution to the problem). Even if the self-driving car can come up with a solution to the problem, it will only act if it’s certain that it has the right answer. If not, given its current configuration, the car will not proceed. Essentially, it just sits there and waits for additional input. Programmers designed it to err on the side of caution.
While it’s an interesting anecdote by itself, this illustrates something much more meaningful. Google are proceeding with a great deal of caution. It’s obviously quite important to the success of their program to get their self-driving cars from testing to production with as few incidents as possible because confidence in the technology is not strong.
The Google Car and Car Safety
Much of the equipment being used on the Google car is already in production and being used on cars all across the country. Car safety features such as lane departure warnings and car-to-car communication are already available on a number of high-end luxury cars and a few top end models of mid-level cars. As a matter of fact, the lane departure system on the car is the same as you’d find on a BMW 7 series.
That sounds great, right? It sounds like this thing is using reliable technology already in existence and that it’s going to be pretty safe. It does sound that way, but read more closely. This car safety technology is equipped on some very costly vehicles. The odds that a self-driving car, loaded with these devices plus an extremely complex operating system and state-of-the-art safety equipment will be cheap or even affordable to the average American is a pipe dream at this point. But that isn’t likely to be a problem. It looks like Google plans the car to be used as a form of public transit—something like a taxi. As a matter of fact, Google has recently invested in the car-sharing service Uber. It doesn’t take a genius to see the potential connection between the ride-sharing service and total car automation. Google already holds a patent for an automated taxi service that would be funded by ad sales. It looks more and more like the Google Car is intended to be a public transit vehicle and less like it’s the next luxury car offering.
For more info on the Google car and a detailed report on its progress to date, we recommend reading this article from The Atlantic Cities.