In the future, new cars might include an appealing sticker: This car is rated for zero fatalities.
Over the next 10 to 20 years, car companies will rely increasingly on computer simulations and virtual engineering to build safer cars and help reduce fatalities. With magnesium and carbon-fiber parts in strategic locations, active safety systems that slow the car as it follows curves in the road, and vehicle-to-vehicle communication that warns you about approaching traffic, future cars will be much safer to drive.
Volvo Car Corp., for instance, has launched a program called Vision 2020, which states, “By 2020, nobody shall be seriously injured or killed in a new Volvo.” It includes not just new protective measures in the car, but technology for communicating dangers to and from the car. Other car companies have similar, less formalized programs.
Ford’s futuristic features on 2011 Explorer
As ambitious as it seems, the zero-fatality goal is achievable, according to Ed Kim, an analyst at automotive research firm AutoPacific Inc. in Tunstin, Calif. In the next 10 years, there will be a confluence of safety technologies — such as road-sign recognition, pedestrian detection and autonomous car controls — that lead to safer cars, says Kim.
Getting there will require carmakers to develop and strengthen an underlying technological infrastructure that can support intense crash simulations using tens of millions of data points, as well as a robust communications network that sends out safety signals between cars. The result will be the deployment of new active safety methods that can predict a crash and drive you out of a jam — literally.
Advanced crash-test simulation
Crash simulation has changed in recent years, says Majeed Bhatti, a car safety engineer at General Motors Co. In the past, engineers designed physical prototypes and ran them into a barrier, then analyzed the results on a computer. Today, the physical prototype is just the last piece of the puzzle. Bhatti says computer simulations are now used as the primary test method.
High-performance computing advancements have enabled GM to move to an interactive design process, according to Bhatti. “We now [simulate] a full vehicle model — that is, bumper to bumper, including interior trim, with [crash test] dummies in it — using as many as four million elements.”
Engineers create test scenarios and send the results back to the designers, who use that information to develop the next round of car models for retesting. “We can see what happens to the trunk, the dummies, and we look at the information as you would do in a physical test. And then we look to see if we need to change body structure, airbags, seat belts — or, in a side impact, the interior trim on door, B-pillar, roof rear headers. All of this would be done after several iterations on a computer,” Bhatti says.
Only after a car design has passed every virtual crash test is a full physical prototype created.
Tools of the trade
Two of the leading software tools for simulating complex crash environments are Altair’s Radioss and ESI Group’s PAM-Crash. GM currently uses LS-Dyna from Livermore Software Technology Corp., originally developed at Lawrence Livermore National Laboratory.
According to Pradeep Srinivasan, a senior technical specialist at Altair, automakers that use Radioss incorporate millions of data elements — such as tire pressure, car weight and road conditions — into a single simulation. One Japanese manufacturer he declined to name used 10 million to 12 million data elements in an especially detailed crash simulation, Srinivasan says.
With such detailed simulation and analysis, it’s not surprising that processing power is an important factor. For many common computer simulations, such as one vehicle crashing into another, carmakers have the supercomputing power they need in-house. Altair has publicly demonstrated that even a complex simulation of a full crash test with 1 million elements can take just five minutes to render using a cluster of Intel Xeon 5500 processors.
American Honda Motor Co. (which includes Honda and Acura cars, as well as Honda motorcycles, motors and power equipment) has more than 3,000 processors dedicated to crash analyses, according to Eric DeHoff, manager and principal engineer for vehicle structure research and computer-aided engineering. “We use high-power computer clusters — load-balancing computers with many processors that share the computational workload — to process many different standardized regulatory and consumer information crash modes,” he says. “We perform structural deformation analyses and occupant injury mode analyses, which require the modeling of restraint system parts like the seat belts, airbags and the actual crash dummies.”
Stuttgart, Germany-based Mercedes-Benz performs thorough simulations on new vehicle designs, says Richard Krüger, manager of safety communications. “We run approximately 5,000 crash simulations with complete car models during the whole development phase [for each car],” he says.
A Honda engineer uses LS-Dyna software to simulate a crash test.
Krüger says Mercedes uses LS-Dyna for the simulation solver for crash tests, Medina from T-Systems for car modeling and Animator4 from GNS for crash visualization to see how the models perform in the rendered environments. “A typical turnaround time for [the complete crash-test simulation] rendering is approximately 15 hours,” he adds
Despite today’s sophisticated testing capabilities, Bhatti says one of the roadblocks to the zero-fatality car is researchers’ understanding of human anatomy and physiology — for example, they don’t know enough about how the brain responds to a head injury to simulate a model that can distinguish between light and severe injuries.
Currently, virtual dummies help measure only displacements, velocities and physical force, Bhatti says. Work is under way — mostly by a group of U.S. and international partners led by the Global Human Body Models Consortium — on building a human model that can measure tissue damage, brain trauma and other accident damage beyond what the typical crash test dummy can show.
David Pulaski, an analyst at market research firm Harris Interactive, related a story about a recent real-world incident for a carmaker he did not name: A thin and light woman was sitting in the back seat of a car involved in an accident, but the car did not deploy an airbag because it was designed to sense a heavier weight and larger size.
If the designers had used more environmental and passenger variables in simulations testing the airbag system, the car could have sensed the woman and deployed the airbag appropriately. This points to a need to expand simulation testing to include more scenarios, which will require even more processing power.
Another major limitation to crash simulations is cost, says Pulaski. Every car company could pour millions of dollars into materials research or virtual human tissue for crash dummies, but consumers would balk at the higher prices car companies would need to charge to offset the costs.
Most of us have accepted the reality that driving cars will sometimes cause fatal injuries. Pulaski says there needs to be another incentive — just as car companies finally started addressing fuel consumption problems when the cost of oil escalated beyond the stratosphere.
Vehicle communication networks
Vehicle-to-vehicle communication is another important step on the road to the zero-fatality car. The more a vehicle knows about other cars (and the roadway), the better it can react and avoid a danger.
Telematics services such as GM’s OnStar and Mercedes’ mbrace today use CDMA cellular and GPS signals to communicate vehicle status, including automatic collision notification, to a central location and provide other services such as roadside assistance and remote door lock or unlock. It’s easy to see how these services could be expanded to allow vehicles to communicate with one another, although neither company has announced specific plans to do so.
Mike Shulman, technical leader for advanced engineering at Ford Motor Co., says his company is moving from passive safety features that protect passengers during a crash to active safety features that can prevent crashes altogether. This means the car will still be made from safe materials and provide airbags and other protections, but it will also actively search for dangers, partly by communicating with other cars and partly by communicating with the road infrastructure, including signs, traffic lights and parking lots.
Pulaski says most automakers have shifted to the active safety approach. For example, just about every car manufacturer now has some form of stabilization control that checks for uneven tire speed and whether the car is at an angle, and keeps the vehicle level to prevent it from overturning. The next step, Pulaski says, is for this stabilization state to be communicated to other cars.
To help spur innovation, the Federal Communications Commission in 2002 approved the use of the Dedicated Short Range Communications (DSRC) 5.9-GHz spectrum for both vehicle-to-vehicle and vehicle-to-infrastructure signals. In 2009, the U.S. Department of Transportation launched the IntelliDrive research program, in which auto manufacturers, in cooperation with federal and state government agencies, are developing standards for the wireless signals and figuring out how to use them in cars. The National Highway Traffic Safety Administration is scheduled to review the program’s recommendations in 2013 and decide whether to approve IntelliDrive technology for deployment.
Today, the Insurance Institute for Highway Safety — the noted U.S. crash-test rating authority sponsored by the insurance industry — rates crash safety based mainly on car-to-barrier collisions. But Ford’s Shulman speculates that the DSRC signals could become part of the IIHS ratings and that future cars could be rated for their ability to communicate with other cars — a glimmer of that possible “zero fatality” rating for cars down the road.
At a low level, these signals would send out a safety state — for example, the car’s speed, the level of brake pressure the driver has applied and steering — to every other car in the vicinity. According to Shulman, this signal would emanate 10 times per second.
A second level of communication involves more detail — the car could send out its current path prediction according to GPS routing, for example, or warn the driver about unsafe traffic conditions reported from other vehicles. One early sign of this was Dash Navigation’s Dash Express GPS system, which debuted in early 2008 and sent traffic information from each Dash owner for other drivers to see. (Research In Motion, which purchased Dash Navigation in 2009, discontinued Dash Express service and support on June 30, 2010, for undisclosed reasons.)
All this could lead to what Shulman calls the smart intersection. Cars would know the status of the next traffic light, the speed of other cars and that, say, there was a semi-trailer truck barreling down the cross street. Shulman says there are other benefits unrelated to safety: Drivers could look up the routes they have taken over several weeks or track their miles per gallon from a computer.
Of course, getting car companies to decide on and conform to an approved standard for car communication may be a challenge. Another challenge is that wireless signals can be unreliable in moving vehicles. For example, Wi-Fi, which is just starting to become available in cars such as the 2011 Ford Edge, requires complex algorithms to make sure it works well in a moving vehicle.
It’s too soon to tell whether the DSRC signals on the 5.9-GHz spectrum will have reliability problems, but there’s no doubt that automakers will need to test and retest the communications systems to ensure uptime and accuracy.
Finally, as AutoPacific’s Kim notes, no amount of vehicle-to-vehicle communication will help when drivers make monumental mistakes, such as driving into a tree.
Collision avoidance systems
To help reduce fatalities, cars’ computers will help drivers prevent crashes in the first place. Adaptive cruise control, which adjusts car speed automatically as you approach another car on the highway; blind-spot warning systems, which use cameras or sensors to detect cars moving up beside you; and lane-departure warning systems, which alert you when you drift out of your lane, are already fairly common. (Read about these features in our earlier story, “Car tech: Taking drivers’ helpers for a spin.”)
Next up are collision avoidance systems that inspect environmental variables such as road conditions, lane markers and your attention level (by measuring steering wheel movements, time elapsed since you started the car, erratic behavior and many other variables) and use advanced algorithms to determine how much you should be braking in a given situation.
Already in use in advanced vehicles such as the Acura RL, the Mercedes S550 and the Volvo S60, these systems send out a radar signal and wait for a response to determine the distance and closing speed of cars and other objects in front of the car. If a collision is likely to occur, the car first warns the driver, then automatically applies the brakes partially or fully, depending on the time to impact.
Similarly, the Volvo S60’s new pedestrian detection system combines radar (to gauge the distance and trajectory of moving objects) with a camera (to determine whether they’re human). If a pedestrian suddenly steps out in front of the car, the system applies the brakes automatically.
The system includes approximately 10,000 stored images of what pedestrians “look like,” according to James Hope, a technical media representative for Volvo Cars of North America in Irvine, Calif.
“The camera compares the stored images to what it’s seeing in front, along with looking for human-type movement — the arms swinging, the shoulders and head, the legs moving — to determine if it’s really seeing a person or something else. The system wouldn’t recognize an animal or a baby stroller, but it would recognize the person pushing the stroller,” Hope says.
These collision avoidance systems require a wealth of data about multiple driving conditions, which allows the computer to factor in every possible variable when deciding upon the appropriate response. As with crash tests, computer simulations have proved key in gathering this data. Volvo tested the S60’s pedestrian detection system over 500,000 km of simulated driving in a virtual world, according to Thomas Broberg, a senior technical adviser for safety at the Volvo Car Safety Center in Gothenburg, Sweden.
To generate the data required for better simulations, Volvo has launched 100 test cars in Europe that are collecting driving-condition data over three years, equipped with video cameras and eye-tracking sensors that will aid further research and simulation. The project will collect about 3 million km of live driving data and will help uncover the results of impacts that are less common than front and rear collisions. Broberg says this data is required if Volvo is to reach its zero-fatality goal.
“There will be a combination of protective enhancements and collision avoidance,” says Broberg. “This will be a groundbreaker for safety, as soon as you get communication between vehicles and between the vehicle and the infrastructure within the next 10 years, if not sooner.” The ultimate goal? Broberg says cars need to get better at performing actions that rely on human interaction today — anticipating what the driver wants to do (say, by evaluating head movements), reading road signs and adjusting speed based on driving conditions. “The car will know its state, that the road is slippery and [the driver] seems tired — there is something coming up in the road, so the car will act evasively,” says Broberg.
Of course, a looming challenge for cars that rely on computers for their safety is that computers are not 100% reliable. Car companies address this problem by creating programs that check and recheck the vehicle state many times in a short period. For example, a Mercedes with blind-spot detection sends out multiple signals to verify that it’s safe to make a lane change; the car also checks the state of the blind-spot detection system itself and warns the driver if it’s not working properly, according to a company representative.
Although we’re unlikely to ever see a car that results in absolutely no fatalities from car crashes, it will be an enormous achievement for the auto industry if it even gets close to that goal. According to the Department of Transportation, 37,000 people are killed annually in crashes on U.S. roadways; if even a fraction of those fatalities could be avoided, we’d be putting technology to its best possible use.
John Brandon is a veteran of the computing industry, having worked as an IT manager for 10 years and a tech journalist for another 10.