For more than a decade, automotive safety has been synonymous with crash management. Airbags, safety cages and crumple zones were less about keeping highways safe, and more about dealing with the huge energy issues during the milliseconds after a crash.
Now, that’s changing. Instead of talking safety cages and crumple zones, automakers and government regulators have begun to espouse a vision of collision avoidance. Instead of air bags and self-tensioning seat belts, they’re talking zero accidents. Zero accidents.
“Now, they’re realizing that the best way to survive a crash is not to have one,” notes Glenn Widmann, chief engineer for advanced integrated safety products at Delphi Corp., an automotive supplier.
Indeed, prevention is the new auto industry hot button, not only in the U.S., but in Europe and Japan, as well. In Sweden, the government is teaming with automaker Saab on “Vision Zero,” a program aimed at wiping out collisions. In Germany, Daimler-Chrysler has discussed an “accident-free driving” vision, while Siemens VDO Automotive has adopted a program called “Zero Accidents.” Japan, meanwhile, is said to be laying the groundwork for smart highways that would eliminate accidents.
Not to be outdone, the U.S. is looking at its own grand vision.On one hand, the U.S. Department of Transportation is studying a concept that would place electronic intelligence at stop signs, traffic lights, and roadsides, in 400-plus urban areas and along 33,000 miles of rural Interstate highways. On the other hand, automotive vendors are spinning a vision of a “safety cocoon” aboard the vehicle, replete with radar, laser, vision systems, and ultrasonic sensors. In essence, both ideas have the same goal: Enable vehicles to “see” their surroundings, provide warnings to drivers and, in extreme emergencies, even commandeer the brakes and steering.
“We believe 95 percent of crashes are preventable,” notes pre-eminent industry consultant, David Cole, director of the Center for Automotive Research. “There will always be crashes that are not preventable, but we can dramatically reduce the number of fatalities on the road today.”
Next Up: The Smart Highway
Increasingly, the concept grabbing the most attention is the new smart highway. To be sure, the automotive world has seen so-called smart highways before. Eight years ago, an industry-government collaboration implemented its own vision of the smart highway, embedding sensors beneath a stretch of pavement along a San Diego freeway and enabling platoons of vehicles to travel in coordinated fashion. The new vision, however, is nothing like that. It requires no infrastructure tear-down and construction. And, in its ultimate embodiment, it would be available to every vehicle on the road, which is why the U.S. Department of Transportation is reportedly investing $50 million over the next five years to build demonstration fleets using the concept.
In essence, the new smart highway is a by-product of the PC world, which is why its development is open to electronics and software companies that don’t have a shred of automotive experience. The key element of the vision is a dedicated short-range communication (DSRC) system, which essentially consists of a radio transceiver board, antenna, and associated wiring. The Federal Communications Commission has already allocated 5.9 GHz as the transmission frequency band for the project, and automotive engineers have begun developing software and hardware for the system. Experts estimate that the on-board bill for such technology wouldn’t exceed $50 per vehicle, since they would only need to add the transceiver board, along with a GPS (global positioning satellite) unit and an extra microcontroller. Highway hardware bills would be similarly minimal.
“You could put one of these radio devices at a traffic light, stop sign, or intersection, and it could broadcast information to approaching traffic,” notes Bob Lange,executive director of structure and safety integration for General Motors Corp. “In the event of a collision threat, it could alert drivers. In extreme cases, it might even be aggressive enough to intervene and apply the brakes or steering.”
The system could accomplish such feats through the use of GPS, which could pinpoint the location of each vehicle near an intersection, and through the use of the transceiver boards, which would enable vehicles to “talk” to each other, as well as to the computer-based roadside systems. In essence, the technical components would form a wireless local area network (LAN) with “headquarters” located at the roadside. By knowing the number, location, and speed of participants in the LAN, the linked system could predict potential problems by drawing on a remote software database connected to the roadside units. A Vehicle Safety Communications Consortium has already launched development of communication test kits for vehicles and roadside modules, using re-engineered 802.11a radio boards and 802.11a antennae mounted on vehicle exteriors.
Engineers say the safety possibilities of such systems are almost limitless. The LAN, for example, could alert vehicles at a traffic light that a driver is about to run a red; it could tell merging expressway driver that they need to speed up or slow down; it could warn lane-changers that someone is driving in their blind spot; it could warn of imminent collisions, approaching emergency vehicles, low bridges, work zones, one-way streets, and even poor road conditions.
The beauty of the system, say its proponents, is that it doesn’t require municipalities to rip up streets. Nor does it run the risk of obsolescence, since its electronics aren’t buried beneath mounds of asphalt. Instead, it capitalizes on the ever-improving nature of electronics technology and leverages the expertise of industry software engineers, who must write the mountain of on-board code that describes scores of potential accident scenarios.
“The exciting thing about this technology is that Moore’s Law applies,” says Lange of GM. “We can leverage the enormous cost reductions in computer-based systems, and we can simultaneously take advantage of the huge increases in compute capacity.”
“If you’re really going to take a major step in accident prevention, you’re going to need to include the highway in the technology equation,” adds Cole of CAR. “The whole idea of the connected car and highway is fantastic stuff, and now it appears to be on the threshold of happening.”
While the Department of Transportation studies the concept of the connected car and highway, tier-one suppliers are hatching a separate plan: to reduce fatalities by placing intelligence on-board the vehicle. Unlike the DRSC system, in which vehicles rely on roadside communications to sense the environment around them, on-board techniques enable vehicles to “see” in all directions, and make informed decisions about their situations.
Some such systems are already available. Electronic stability control (ESC), which combines yaw sensors and antilock brakes as a means of stopping a skid, is now widely recognized as one of the industry most effective methods of reducing fatalities (see sidebar). In June, the Insurance Institute for Highway Safety estimated that across-the-board use of ESC could eliminate 10,000 fatalities annually in the U.S.
Efforts to reach the zero accidents plateau call for more extensive measures, however. Suppliers and automakers are working together on those measures, and have laid the groundwork for a progressive plan that begins with adaptive cruise control, then moves to autonomous lane keeping, and graduates to collision avoidance. All of the technologies require use of advanced sensors. TRW Automotive, for example, markets an adaptive cruise control system that employs a 77-GHz long-range radar sensor. The same company is also working on a lane-keeping system that employs forward-looking video sensors that mount above the rear-view mirror. Such systems could one day work in an active mode, signaling the steering to move a car back toward the center of its lane if it begins to drift.
For automakers, the challenge inherent in such systems is the fusion of massive amounts of data from radar, lasers, video, yaw rate sensors, and GPS units. Video systems, for example, can more effectively determine whether an object in the road is dangerous by working with GPS units and radar. Is the object moving? How far away is it? How fast is the vehicle traveling? All of those questions might need to be answered to determine whether danger is imminent.
“Obviously, as we fuse data, the computational requirements rise,” notes Phil Cunningham, director of product planning for TRW Chassis Systems.
Indeed, engineers face a computing dilemma, not only because of the huge amount of sensor data, but because its complexity. To make sense of it all, automakers want to combine the sensor inputs to form a software model that creates a “holistic” picture of the outside environment.
That’s important, experts say, because if vehicles can ever be expected to steer themselves out of potential crashes, they must first know everything that’s around them. That way, they don’t steer out of one crash and into another, worse one.
For that reason, vehicle intelligence is likely to evolve in a hierarchical pattern. At the top of the hierarchy, very powerful central processors will manage input from powertrain, chassis, infotainment, and interior power systems. Those central processors typically will employ a cluster of four or five PowerPC-type microprocessors, which will examine “pre-conditioned” data from sub-systems.
“It might take input from the navigation system and, knowing a curve is ahead, it could command the transmission to downshift,” says Peter Schulmeyer, director of strategy and marketing for Freescale Semiconductor’s Transportation and Standard Products Group. “It doesn’t replace the navigation system and it doesn’t replace the transmission controller. But it adds a layer of intelligence above them.”
In the hierarchy, a layer of sub-systems would operate the ABS, airbags, engine, and transmission, taking instructions from the central processors. Beneath that layer, microcontroller-endowed sensors -- for the airbags, vision systems, and navigation systems -- would gather the data.
“It would require very different processing than we have today,” Schulmeyer explains. “A lot of the processing would take place in the radar system or the vision system. But the decisions would be made by the central processors.”
A Question of Cost
The question of whether automakers will opt for smart highways or on-board intelligence is one that will most likely be answered by cost. Many observers doubt that low-end vehicles will ever be able to afford large arrays of costly laser, vision, or radar sensors. Conversely, most observers also believe that many medium- to high-end vehicles will feature such technology.
“To get 360-degree coverage around the vehicle, you’re going to need 10, 11 or 12 sensors and additional CPUs on-board the vehicle to manage all the input,” notes Lange of GM. “As effective as that sensor fusion can be, it’s difficult to envision a circumstance where you could put multiple ten-dollar sensors like those on a low-end Chevrolet Aveo.”
Lange adds, however, that automakers are likely to adopt the technology for higher-end vehicles. “It’s going to happen,” he says. “We’re doing it. Ford, DaimlerChrysler, and Toyota are all working on those types of technology, too.”
Using either method, experts argue that a large percentage of today’s fatalities are preventable. That’s why NHTSA is expected to raise the bar this fall with a Notice of Proposed Rulemaking mandating stability control-type technologies in the U.S. during the next three years. Meanwhile, a consortium of 12 manufacturers are working on safety-related “use cases,” or software models, for DSRC-based smart highways.
“It’s unrealistic to expect to achieve zero fatalities, even in the next 20 years,” notes Adrian Lund, president of the Insurance Institute for Highway Safety. “But it’s important to recognize that all fatalities can one day be preventable. We need to understand how they occur, and then we need to work on preventing them.”
Three Ways to Reduce Fatalities … Now
In the long term, advanced sensors and smart highways are the keys to zero fatalities. But experts say that existing technologies could have a huge impact on the reduction of the 43,000 highway fatalities that occur in the U.S. every year. Here are three such technologies:
Electronic stability control. It’s been around for a decade, and it’s popular in Europe, especially in Germany. But it’s a standard feature in only about 35 percent of the vehicles sold in the U.S. The Insurance Institute of Highway safety estimates the technology could save 10,000 lives per year here.
Alcohol ignition interlocks. Although the Insurance Institute for Highway Safety certainly doesn’t recommend interlocks for all drivers, it has said that 8,000 lives could be saved annually by preventing alcohol-impaired drivers from getting behind the wheel.
Seat belts. Surveys have shown that 82 percent of vehicle occupants use seat belts. Boost that number to 100 percent, and fatalities could drop by another 5,800 annually.
The Drive Toward Stability Control You may not have electronic stability control on your vehicle right now, but you will soon. General Motors Corp. has announced that all of its new vehicles will incorporate its brand of stability control, called StabiliTrak, by 2010. Moreover, NHTSA is expected to issue a Notice of Proposed Rulemaking this fall related to stability control. Ultimately, a mandate could put the feature on all vehicles sold in the U.S. sometime in the next three to four years. The advantage of stability control is its ability to make neophyte drivers look like experts.ESC yaw rate sensors recognize skids, then alert ABS units. By applying the brakes at the proper wheels, ABS modules stop the yaw. “Keeping the vehicle on the road is paramount to reducing tripped rollovers,” notes Paul Mercurio, marketing manager for Bosch Automotive’s Chassis Division, which makes stability control systems. “The advantage of electronic stability control is it keeps drivers on the road.” On most vehicles, ESC adds between $300 and $800 to the price of the vehicle. Yaw rate sensors and steering angle sensors need to be incorporated, and ABS units must be upgraded. Most ABS units require four extra hydraulic valves for fluid flow in and out of the pump. Electronic controllers and software must also be added. In its effect on safety, ABS has exceeded automakers’ wildest dreams. A study released by the Insurance Institute for Highway Safety earlier this year indicated that stability control could save 10,000 lives per year. “Every vehicle is susceptible to tripped rollovers,” Mercurio says. “It doesn’t matter if you’re driving a Ferrari or an SUV. Once it gets tripped, you’re going over.” Sensors for the Next Generation To meet the needs of next-generation safety systems, automotive suppliers are rolling out a new breed of sensors, ranging from lasers to radar to gyroscopes.
TRW’s 77-GHz radar sensor captures data up to 200 meters away. It serves in the company’s adaptive cruse control technology, which automatically applies the brakes when a slower vehicle appears in front.
Bosch Automotive’s micromechanically-fabricated DRS MM3 gyroscopic sensor allows detection of angular velocity for electronic stability control, rollover prevention, and active steering.
Sick Optic’s LMS 291 laser sensor served on the winning autonomous vehicle in DARPA’s $2 million, winner-take-all, Grand Challenge race. It “paints” a 3-D picture of the terrain in front of a vehicle.
How Smart Highways Will Help Drawing on a huge database of potential accident scenarios, smart highways could alert drivers to a multitude of potential dangers. Here are a few of those:
Traffic signal violation warning: Predicts if drivers will run a red light.
Left-turn assistant: Helps drivers to decide when – and when not to – turn left into oncoming traffic.
Intersection collision warning: Warns drivers if a collision with intersecting traffic is imminent. In its ultimate embodiment, the system could take control of the brakes and prevent an accident.
Approaching emergency vehicle warning: Broadcasts a message including speed, location, lane information and intended path of approaching emergency vehicles.
Low bridge warning: Warns commercial vehicles of impending low bridges.
One of the biggest walls in embedded software development is the integration of low-level drivers with higher-level middleware and application code, but silicon vendors are stepping up to bring it down.
Focus on Fundamentals consists of 45-minute on-line classes that cover a host of technologies. You learn without leaving the comfort of your desk. All classes are taught by subject-matter experts and all are archived. So if you can't attend live, attend at your convenience.