Collision avoidance and intelligent diagnostics systems aside, the ultimate smart car is one that drives itself, with little or no input from a human driver. This type of self-driving car will use sensors, computers, and other smart technologies to sense where it is in relation to other vehicles (and the road), and navigate to its destination according to preset coordinates.
For a car to operate completely autonomously, without any assistance from you behind the steering wheel, a number of different technologies have to be employed. These include:
- 360-degree cameras, to view around all sides of the car
- Adaptive cruise control, to regulate speed in traffic
- Emergency brake and steering assistance, to avoid collisions
- GPS, to determine precise location and navigate routes
- Radar and LIDAR, to sense the distance between your car and other cars or objects
- Stereo cameras, to outline and identify pedestrians and bicyclists as different objects than other cars
By combining collision avoidance systems with adaptive cruise control, a car will be capable of driving down the road without hitting anything in its way. The exact route is specified via GPS mapping and navigation, of course. Program in your destination (or just say it; you might as well add voice control to the mix) and the car’s computer calculates the best route there. The automatic systems then take over and drive the route.
The good news is that most of the automatic systems needed for a self-driving car are in existence today. Not surprisingly, most automobile manufacturers have big plans for their own autonomous vehicles. For example, Audi plans to market, in the near future, vehicles that can steer, accelerate, and brake, all by themselves. Cadillac plans to introduce select models with autonomous lane keeping, speed control, and brake control. Mercedes plans to introduce the Highway Pilot system that enables hands-free highway driving with autonomous overtaking of other vehicles. And Toyota is working on what it calls near-autonomous vehicles with Automated Highway Assist, Lane Trace Control, and Cooperative/Adaptive Cruise Control.
If you want to see the future of self-driving vehicles, however, you have to look beyond the traditional automobile manufacturers to the world of Silicon Valley. There, Google (yes, that Google) is at the forefront of autonomous driving technology.
Why Google wants to develop self-driving cars is another story, rooted in its drive to create street view photographs of seemingly the entire world for its Google Maps service — and using self-driving cars to take those pictures. Officially, Google says that it’s goal is “improving road safety and transforming mobility for millions of people.” Whatever the motivation, the company is making great strides in the field.
Google started out by retrofitting a variety of stock vehicles — Audi TT’s, Lexus RS450s, and Toyota Priuses — with autonomous driving hardware and software. With that experience under its corporate belt, Google then took the next logical step and built its own prototype self-driving car. The company plans to build 200 or so of these robotic driving machines, all set to tool around California for the next year or so.
The Google car (it has no formal name) is not a scion of automotive fashion. It’s kind of cute, actually, especially with the perceived smiley face in the front. To me, it looks a lot like those Little Tykes plastic toy cars that my grandkids used to roll around in.
It’s not a toy, however. The Google car is a battery-powered electric vehicle, capable of a maximum speed of 25 mph. It has a stop/start button, but no steering wheel or pedals. There’s room for two inside, although both are passengers, not drivers.
Inside Google’s car is a variety of sensors and technologies to detect the vehicle’s surroundings. We’re talking stereo cameras, 360-degree cameras, LIDAR, radar, and even sonar devices. All these devices are necessary because they look at things differently; they all have different ranges and fields of view, thus serving a particular purpose in the grand plan.
For example, stereo cameras — two cameras with a small separation between them — are mounted around the exterior of the car. They’re used to create overlapping fields of views that can track an object’s distance in real time. These cameras have a 50-degree field of view, but are only accurate up to about 30 meters.
More accurate distance is measured by the car’s LIDAR system, mounted on the top of the car. This system is powered by a Velodyne 64-beam laser that can rotate 360 degrees and take up to 1.3 million readings per second. The LIDAR system is accurate up to 100 meters, which makes it ideal for generating a real-time map of the car’s surroundings.
Radar systems are built into the car’s front and back bumpers. These systems are used to warn of impending impacts and tell the brakes to activate when necessary.
Interestingly, the radar systems are paired with sonar systems. That’s because the two technologies are best for different distances. Radar works up to 200 meters away, while sonar is good for distances of 6 meters or less. Combine the two and you get accuracy at all distances.
These sensors generate up to 1GB of data every second. This data is used to build a map of the car’s immediate surroundings, which enables the car to stay in the correct lane and avoid obstacles.
Google’s self-driving car has a long way to go before it’s ready for the consumer market. (It has to drive faster and carry more passengers, to begin with.) But it sheds some light on where automotive technology is going. It won’t be that long until your new car not only can avoid collisions and park itself, but can also drive you to where you want to go. Just hop in, tell it your destination, then settle back and take a nap. Tomorrow’s smart car will do the driving for you!