see. think. act.

Technology

#VisionZero #AutonomousEverything

Learning from Crashes

Min Reading Time
Tags: ZeroAccidents, Safety
Vehicle accidents can be nerve-racking and may have serious consequences. But we can learn from them. Those who understand how they arise will be able to assist in reducing them in the future. That motivates ZF accident researchers.
Achim Neuwirth, December 10, 2019
author_image
Achim Neuwirth has been writing for ZF since 2011. He has specialized in writing texts about all kinds of car-related topics: from vehicles to the technology behind them, to driving and traffic.
Too many drivers have tried to get around a traffic jam on the freeway. Now there is stop-and-go traffic on the detour that bypasses the jam, and the long line of cars ends in a bend in the road that drivers cannot see around. The first impatient driver shears off from the line and wants to make a U-turn in order to look for what they believe will be a better alternative route. They overlook the passenger car approaching in the oncoming lane. Will this driver stop in time, or will they T-bone the side door at high speed? At the same time, a van is approaching the traffic jam from behind. The stressed delivery driver constantly has one eye on his cell phone. Will he see the hazard lights flashing in front of him in time?

Every driver is familiar with risky situations. Head-on, side-impact, and rear-end collisions in which the occupants are seriously injured or killed happen too often. It is hard to see anything positive about such moments, but at least you can learn from them. Crashes and dangerous situations – near misses in particular – provide a lot of information. Analysis of this data can help prevent collisions or reduce their consequences in the future. At ZF, this research is studied, and lessons learned are embedded in the DNA of the company’s active, passive, and integrated safety and security systems.

Recognizing dangers

Recognizing dangers

It is about observing where real accidents and risky situations occur and finding out what chain of events brought them about in each case. The consequences of the crashes are assessed, including critical injuries suffered by people and damage caused to vehicles. They can also forecast what the focal points for accidents and hazards will be in the future. This is one of the first tasks of the accident research being conducted in ZF Advanced Engineering: “If we discover white spots on the safety map, we provide advice, for example whether we can cope with these new areas using existing systems or whether a new solution should be developed,” says Tobias Hänsel, Project Coordinator in Advanced Engineering. “In principle, it is an intensive search for methods for achieving better or new safety solutions that bring ZF closer to Vision Zero – the goal of zero accidents.” Hänsel is currently using computers to see how these system ideas perform in a virtual accident. To create data for this simulation, the team made use of tools including GIDAS. GIDAS stands for German In-Depth Accident Study, an accident database held in high regard around the world (see end of article).
3500
individual items of information are recorded in the GIDAS database for each investigated road accident, representing an unparalleled volume of detail. This is what accident researchers use for their analyses, forecasts, and improvements – in the automotive industry and, for example, in legislation, traffic planning, and emergency services.

Assessing protective effects

Assessing protective effects

However, sometimes there is no need to conduct in-depth research in order to identify problems or areas where action is needed. For example, accidents involving turning trucks occur too often. And side-impact crashes involving passenger cars can have consequences whose severity is above the average. “Therefore, our team often starts with the second step, the project-related analysis, into the advanced engineering task,” says Hänsel.

This is what happened with the pre-crash system with external side airbag, which ZF presented in June 2019: GIDAS provided the database to set up virtual side-impact scenarios. In these scenarios, the experts simulated a typical mid-size car with a set of crash test dummies inside. In the digital model, the vehicle was equipped with the new pre-crash side-airbag system. The team then simulated scenarios in a wide variety of ways. They modified the speeds, impact angles, collision types, and type of opposing vehicles involved and also adjusted the technology itself on the computer. “We examined a pool of more than 28,000 scenarios in total,” says Hänsel. When it was all over, the results were documented with documented figures. “The pre-crash side airbag allows the consequences of the accident for occupants to be reduced by up to 40 percent, and the penetration of the opposing vehicle into the car body by up to 30 percent,” says the analyst. In the situation described at the beginning of this article, the impatient driver would an improved chance of survival, even if the oncoming passenger car was to strike their vehicle at 65 kilometers per hour.

At the same time, a system named Automated Front Collision Avoidance would make a side-impact and rear-end collision involving the approaching van less likely to occur. The integrated ZF safety and security system is capable of performing emergency braking and evasive maneuvers for the driver. ZF has already rolled out similar solutions for heavy trucks.
The pre-crash side airbag and Automated Front Collision Avoidance also make it clear: At ZF, accident analysts are right in the thick of the advanced engineering process. They evaluate new driver assistance and occupant safety systems on a virtual basis. Only then do such systems advance to testing in the real world, to volume production development, and ultimately to the market.
28000
was the number of different crash scenarios that ZF accident analysts simulated virtually just for the pre-crash side airbag. Such simulations make it possible to help digitally evaluate how well new safety and security systems can work.

International implementation

International implementation

The ZF crash researchers are also involved in advising safety committees throughout the world. They do this with the goal of inspiring others to create additional accident databases around the globe. This first comes down to more such databases that resemble GIDAS, such as the established Fatality Analysis Reporting System (FARS) or the Crash Report Sampling System (CRSS) in the USA, or ones like the China In-Depth Accident Study (CIDAS). Among other things, they also form an information basis for further development of seat belts and airbags – that is, for systems that primarily work when a crash cannot be avoided.

But before the crash occurs, modern assistance systems can support the driver in critical driving situations and help prevent accidents, such as by automatically applying the brakes ahead of a potential collision or by taking evasive maneuvers. Through the analysis of near misses, crash researchers can better understand how accidents occur and how assistance systems can be designed to be less prone to false alarms. Naturalistic Driving Studies (NDS), carried out by research organizations such as the Virginia Tech Transportation Institute (VTTI) in the USA, are regarded as foremost among such analyses.
Volunteers have camera and sensor systems installed on their vehicles. They record real, ordinary driving scenarios and critical situations and make them available for evaluation. Above all, the accident researchers’ ultimate wish is that the information in all the collections will be able to be directly compared. “There is still a long way to go before we get there,” Hänsel points out.

As with the databases, it is also the case that accident scenarios, vehicle types, and crash test requirements differ. To respond to these issues, ZF has well-trained safety experts in all regions of the world. “We are continuously expanding this network of international specialists,” Hänsel emphasizes. “One example is the cooperation between our Chinese colleagues and the experts working on the CIDAS.”

Databases that are more and more uniform, a growing number of proven safety experts, technology and simulations that are getting better and better: These are the key instruments of leverage for bringing tailored systems onto the roads more rapidly. “It is estimated that it will still take about nine years for a new solution to be established across the board,” states Hänsel. “The faster we get to this point, the sooner advanced safety technology can demonstrate its ability to significantly reduce the frequency and consequences of accidents.”

Video: The ZF experts examine a wide variety of scenarios and evaluate which sensors and systems can be effective in such a situation and can help to prevent a possible accident. For example: A vehicle approaches at a constant speed and does not recognize the pedestrian. A sensor detects the pedestrian and determines that a collision will occur. The vehicle can automatically initiate a braking maneuver helping to prevent this accident.

Data for protection: GIDAS

Data for protection: GIDAS

German In-Depth Accident Study (GIDAS) is one of the world’s largest and most sought-after road accident research projects. Each year, it records some 2,000 real crashes involving personal injury in the metropolitan areas of Hanover and Dresden. When there is a crash, a survey team of technicians and physicians go to the scene and gather up to 3,500 individual items of anonymous information and images. Things like the sequence of events and type of accident, the directions of travel, and the various positions and technologies fitted to the vehicles are documented in a very precise and in-depth manner. In addition, medical information of the injured is collected. The sequence of the rescue chain, how the scene of the accident was secured, and what the traffic flow was like at the time of the accident are also described.
As a partnership between Germany’s Federal Highway Research Institute and Germany’s Automotive Research Association, GIDAS wants to achieve one thing above all else: Legislative bodies, automotive manufacturers, and suppliers are to get the opportunity to very closely monitor what is currently happening during accidents and to identify trends. This allows technology to be improved and helps make road transportation safer for everyone. With the same goal in mind, ZF – as one of the founding member - makes intensive use of the GIDAS data.