No, it’s not science fiction – autonomous vehicles are cars that can sense the environment and operate with no humans involved and they’re up and running now. Not only do you not have to drive the car, but these cars can also run without you inside of it at all. Autonomous vehicles can do everything that a human driver with years behind the wheel can do, and go anywhere your current car can go.
These cars are based on a sophisticated computer that processes everything in the environment, plots out a path, then sends instructions to the vehicle’s actuators. An actuator is simply a machine part that is responsible for moving and controlling a mechanism (like a car!).
The actuator controls acceleration, steering, and braking. Rules coded into the computer, as well as the computer’s avoidance algorithms, predictive modeling and object recognition all help the computer software avoid accidents and obstacles and follow traffic rules.
So how do autonomous vehicles work?
These vehicles use cameras, lidars, machine learning, maps, and radars to drive on their own. They will create a map of the surrounding environment based on the sensors found in different parts of the car. For example, ultrasonic sensors in the wheels can detect other vehicles when parking and avoid curbs.
The tools that self-driving cars use:
Cameras – cameras are used to identify speed signs, lane divisions on the highway, and traffic lights. Developers believe that as machine vision improves, cameras can be used to identify everything they see and navigate accordingly.
Lidars – lidar is light detection and ranging, and you’ll see it on top of an autonomous vehicle. It works by shooting out millions of laser beams per second, then measures how long they take to bounce back. The car’s computer measures how long they take to bounce back and uses this information to construct a 3D map more precise than radar.
Machine Learning – this is artificial intelligence, which will train the computers in the car to identify pedestrians or identify lane lines. The computers learn by seeing millions of examples of every possible scenario. That is, every scenario that can be predicted – it wouldn’t be possible to write our computer rule for everything that might happen. So the cars are able to learn from experience and use artificial intelligence to navigate on their own.
Maps – the autonomous vehicle manufacturer will use cameras and lidars to map the area where the car will operate in extreme detail. The computer running the car can take this as a reference to verify its sensor readings. Cars need to know their own location, right down to the centimeter.
Radars – radars bounce radio waves around to get a feel for the surroundings and do particularly well spotting other cars. They can work in the fog, rain, or snow and are reliable and cheap.
Challenges Autonomous Cars Will Have to Overcome
Lidar is expensive, and developers are still trying to find the right balance between resolution and range. For example, if two autonomous cars driving on the same road, would their lidar signals interfere?
When autonomous cars drive on a layer of snow, lane dividers can’t be identified. Can cameras and sensors keep track of lane dividers if they are obscured by environmental elements like water, debris, oil, or ice?
Benefits of Autonomous Cars
As scary as the idea of a self-driving car may be, autonomous vehicles present a lot of benefits to the public. First, you can imagine how convenient a car that could drive itself might be and how this might improve the quality of your life. For example, if you had vision problems or were physically disabled, you could still get around town. What if your kid forgot lunch at school? Just send the car to her school to bring her lunch. Is your cat sick? Send her to the vet with your self-driving car.
Autonomous cars would also be highly beneficial for the environment, drastically lowering CO2 emissions.
Additionally, by 2050, autonomous cars:
- It would reduce traffic congestion, as a result of 30% fewer vehicles being on the road.
- It would slash transportation costs by 40% when you consider vehicles, infrastructure, and fuel.
- Improve livability and make it easier to walk around
- Reduce CO2 emissions by 80% worldwide
Although cars run by computers present exciting prospects for humans, developers still have a lot of work to do to perfect their models. When do you think autonomous vehicles will commonly be seen on the road?