Robocars: Are Self-Driving Cars Next?
By Salvatore Salamone
Posted Aug 5, 2008, at 1 p.m.
When it comes to cars that drive themselves, science fiction may have over-promised, as there is a long, rich history of imagination to catch up with. However, progress is being made to a point where successful demonstrations continue to push the envelope.
The recent Toyota Grand Prix race in Long Beach included three driverless cars. Such so-called robot cars might one day be capable of traveling from point A to point B without a driver. But the real value of the combined navigation, crash-avoidance and artificial intelligence technology at the heart of these cars is the improved safety when a human is behind the wheel. For instance, the technology might be able to step in when a driver is incapacitated and bring the car safely to a stop (or perhaps a hospital).
Elements of a Driverless Car
Self-navigating vehicles require navigation systems, lane-departure warning systems and crash-avoidance systems. Many of these are already available:
- Voice-guided GPS navigation is built into many new car models, and it’s available as add-on hardware and as a third-party service in other models.
- Lane-departure warning systems are available today on many premium cars, including models from BMW, Buick, Cadillac, Infiniti, Lexus, Mercedes-Benz and Volvo.
- Active Cruise Control (ACC), also known as adaptive or intelligent cruise control, uses radar or a light beam and intelligent software to adjust the car’s speed and maintain a safe distance between vehicles when driving in traffic. Such systems are now available in select models from Acura, Audi, BMW, Cadillac, Lexus, Mercedes and Volvo.
- Intelligent software that aggregates the information from various systems is being developed by government and industry initiatives, such as the U.S. Department of Transportation’s Integrated Vehicle-Based Safety Systems (IVBSS) program.
If you combine all of these technologies, there is still one key component that must be developed: artificial intelligence (AI).
AI software is the missing link that can pull together information from all of these systems and make driving decisions as the car travels from point A to point B.
For instance, the goal of IVBSS is to integrate information from numerous systems and present drivers with alerts to avoid crashes, stay in their lanes and slow down to take an upcoming curve safely. Such a system is not self-navigating, but its information could be a critical part of the data used by a self-navigating car’s AI application. (Other technology components that could help are cameras or image recognition systems and active elements in the road that pass along information.)
The automotive and trucking industries are not the only ones developing technology for self-navigating vehicles. Mining, space and military applications have been interested for years.
In mining efforts, the emphasis is on developing vehicles that can map out underground mines. Space applications have focused on developing mobile vehicles that can operate on planets, moons and asteroids. And military efforts include the use of self-navigating vehicles to deliver supplies to troops, address dangerous situations (bomb disposal, for example) and scout out enemy territory.
Developments With Self-Driving Cars
While today’s self-navigation technologies do not operate as brilliantly as those of fictional movie and TV cars, recent demonstrations are fairly impressive.
The Defense Advanced Research Projects Agency (DARPA) Urban Challenge held last November required teams to build an autonomous vehicle capable of driving in traffic and performing complex maneuvers such as merging, passing, parking and negotiating intersections. DARPA claimed the event was truly groundbreaking, “as the first time autonomous vehicles interacted with both manned and unmanned vehicle traffic in an urban environment.”
Thirty-five teams were invited to a National Qualifying Event (NQE) that required the cars to pass three tests. First, the cars had to merge safely into and out of two-way traffic in a tight course. Second, the cars had to stay within a lane as they traversed a 2.8-mile course that included a section, called “the Gauntlet,” where they had to maneuver through a series of parallel-parked cars and road obstacles. The third test comprised a series of four-way stop intersections in a traffic-intensive environment.
Eleven teams passed the NQE and, a few days later, participated in the final event. The cars had to navigate a course and visit a number of checkpoints, in a particular sequence, while avoiding obstacles, other cars (about 30 manned vehicles driven by professional drivers) and each other. Six cars finished the course.
Looking to the Future
As self-navigating technologies are integrated into vehicles, many initial applications will probably look to combine auto-pilot with human driving.
Some trucking firms would like to use driverless trucks to make long-distance runs. But more likely, the technology would help truck drivers on long runs with something akin to an enhanced cruise control and with speed and proximity alerts.
Similarly, the artificial intelligence associated with self-navigating cars might be able to detect when a driver is medically impaired (by monitoring vital signs and hand-eye activity) and is deviating from a lane or not responding to traffic conditions (e.g., continuing at a high speed into a very congested traffic situation). Having noted these conditions, the car might take over control and guide the vehicle to a safe stop at the side of the road.
Integration of self-navigating cars with other technology holds even more promise. For instance, the combination of GPS and mapping technology could be leveraged so that cars can be instructed to drive to the nearest hotel, restaurant or hospital.
While it may take many years for truly self-navigating cars to hit the market, a number of key technologies are already available, and the rest are being developed in real-world tests like the DARPA Urban Challenge each year. But the progress being made today indicates the technologies for robocars will likely be part of many cars within the next 10 to 20 years.
Bio: Salvatore Salamone is a senior editor at Ziff Davis Enterprise. He has more than 20 years of experience writing about science and technology for major industry trade magazines and is the author of three business technology books.
Keeping Solar Accessories Powered Up
Permalink: http://www.progressive.com/auto-tech/are-self-driving-cars-next.aspx
