Earlier this year, Google began testing its fleet of purpose-built driverless cars on public roads. It wasn’t so long ago that the idea of an autonomous car, controlled by a computer instead of a driver, seemed like a fantasy. However, over the past few years, driverless technology has progressed at a rapid pace, and many major manufacturers expect to bring a driverless car to the market by 2020. The idea for a driverless car has existed for decades, but as lawmakers and manufacturers tackle the remaining issues, we may soon reap the benefits of this promising technology.
How Do Driverless Cars Work?
Driverless cars are packed with sensors, such as GPS receivers, inertial navigation systems, high-resolution video cameras, laser rangefinders and LIDAR systems that enable the car to calculate its position and understand its immediate environment. They also take advantage of the latest image processing software, allowing the cars to understand traffic signals and recognize the difference between a pedestrian and a fire hydrant.
However, merely being able to sense the world around it is not enough to create a driverless car. These vehicles must be able to react autonomously to real-world conditions, requiring advanced software to control the car, plan routes and make decisions about common driving situations. Driverless cars cope with these demands in many ways; for example, most are equipped with detailed maps, accurate to within an inch, and sophisticated route-planning software. They’re also equipped with limited artificial intelligence systems, allowing the vehicle to respond appropriately to obstacles such as road debris, pedestrians and other vehicles.
All of these technologies work together to drive efficiently and safely. For example, the vehicle regularly checks its built-in map against data from its sensors and updates the map. Data from each sensor is checked against complementary sensors to prevent an equipment failure from causing an accident. Once the car’s internal map is generated, it calculates an optimal route, taking into account both its final destination and its immediate surroundings. The vehicle updates both its route and its internal map multiple times each second, allowing the vehicle to adapt to changing situations.
The Development of Driverless Cars
The idea for driverless cars has been around almost as long as cars themselves. In the 1920s, Houdina Radio Control developed the Linrrican Wonder, a radio-controlled vehicle that accepted radio commands from another vehicle. For the 1939 World’s Fair, Norman Bel Geddes presented the idea of a fleet of radio-controlled vehicles to an awestruck public in his Futurama exhibit. Bel Geddes envisioned his vehicles taking commands from wired roads.
Although several test projects were actually built based on Bel Geddes’ ideas, these early driverless cars weren’t autonomous vehicles. They could follow a predetermined track, but they couldn’t respond to unexpected situations or deviate from their instructions. Intelligent driving still required a human operator, even if the operator wasn’t in the vehicle.
The 1980s saw the first autonomous and semi-autonomous vehicles as universities, auto manufacturers and government agencies began to pour serious resources into the development of driverless vehicles. Many important milestones were achieved during the ’80s and ’90s, such as the first driverless trip along a regular public street, autonomous navigation through off-road conditions and semi-autonomous driving in regular traffic.
However, over the past few years, researchers have made rapid progress towards deploying viable driverless vehicles. Google’s fleet of driverless vehicles have logged hundreds of thousands of miles in regular, everyday traffic without major incident. Most of the major automakers, including GM, Ford, Volkswagen, Mercedes-Benz and Nissan, have spent millions of dollars developing driverless vehicle systems.
Driverless Vehicles Are Already Here
Although commercially viable autonomous vehicles aren’t yet a fabric of everyday life, driverless vehicles have in fact already become a reality. In 2010, Italy’s VisLab, a major player in artificial vision research, sent a small fleet of driverless vehicles on a 9,900-mile journey from Italy to China. The Chevrolet EN-V series of prototype vehicles, first unveiled in 2011, features autonomous operation, vehicle-to-vehicle communications and smartphone integration in the real urban environments of Shanghai.
In fact, according to Jason Lancaster, the editor of AccurateAutoAdvice.com, we’ve already reached the first major milestone in the integration of driverless cars into the mainstream auto market. This first milestone is characterized by limited, self-contained autonomous technologies, such as adaptive cruise control, parking assist and lane keeping. These technologies are capable of exerting some control over the vehicle, but they’re largely designed to assist human drivers. Even the more comprehensive technologies, like Volkswagen’s Temporary Auto Pilot and Toyota’s Intelligent Transport System, are more concerned with helping the driver reduce accidents than at taking control from the driver.
As these technologies mature and become more robust, we’ll pass through the next two stages of Lancaster’s predicted transition. In the second stage, autopilot technologies will communicate with their fellow vehicles, creating a map shared by all the vehicles in the area. Humans will still retain ultimate control of the vehicle, but autonomous technologies will take over most of the tedious aspects of driving. Finally, in the third stage of the transition, high-resolution sensors will combine with vehicle-to-vehicle communication and autonomous technologies to deliver a completely driverless experience.
Challenges of Driverless Vehicles
Before they can become a common element on public streets, driverless vehicles will have to overcome several challenges, including:
- Weather. Most of the advanced sensors used in driverless vehicles work well in good weather, but rain and snow can render some of the sensors useless.
- Safety. Driverless vehicles will have to recognize a dangerous situation and calculate the best way to either minimize the danger or avoid the situation entirely.
- Human drivers. The transition to a fully autonomous fleet won’t happen immediately, so driverless vehicles will have to be able to recognize when a vehicle is driven by a human and react appropriately.
- Loss of control. Although not an issue with the cars themselves, automakers will have to convince drivers to allow the car to drive itself and to give up their control of the vehicle.
- Legal issues. Currently, only a handful of states allow autonomous vehicles to use public roads. Additionally, liability issues are a concern: If no one is driving the car, who is at fault in case of an accident?
Society will have to solve several other concerns as driverless cars become a reality. For example, professional drivers are likely to find themselves replaced with self-driving vehicles. Hackers might try to compromise driverless cars. The initial driverless cars aren’t likely to be cheap, and many of the benefits of driverless cars won’t be real until most cars are driverless. Finally, if governments mandate that all cars use autonomous technologies, drivers who legitimately enjoy driving will have fewer opportunities to drive.
Benefits of Driverless Cars
Despite the challenges, driverless cars offer several tantalizing benefits. Driverless cars don’t fall asleep while driving, nor do they text their friends at stop lights. Distracted or intoxicated driving would become a thing of the past, and commuting would turn into personal time.
Driverless cars could also dramatically reduce traffic congestion.
Autonomous technologies would allow cars to follow each other closely without impacting safety, and these vehicles wouldn’t become morally outraged at the thought of allowing another car to merge into traffic.
However, the most attractive benefit of driverless cars is the possibility of drastically reducing accidents and traffic fatalities. With their advanced sensor suites and their ability to update their internal maps thousands of times each second, autonomous cars would have few, if any blind spots. They could notice potential accidents as they began and react by the millisecond, plotting alternate escape routes or applying the brakes as necessary to prevent the accident from happening.
Although there are technical, legal and cultural roadblocks to overcome before autonomous cars become a reality, the idea of a fast-moving, accident-free future is temptingly close. These cars will be here sooner than we think. Google expects to release its first commercial autonomous vehicle by 2018, and most manufacturers will have some form of autonomous vehicle available by 2020.
What do you think of a future without drivers? Are you excited to be able to watch movies on your smartphone while you commute, or would you rather stay in control of your own vehicle?