Effects on the simulated human bodies in the software are studied to improve devices such as airbags.
Toyota Motor Corp. said it is including child-size models for the first time in the latest update of its virtual crash test dummy software, which is sold globally to automakers, suppliers and researchers.
Although Toyota began research on child models in the first version of the Total Human Model for Safety software in 2000, a company spokesman said, there wasn’t enough research completed to include the models in the software.
Toyota’s software simulates bodily injuries during vehicle crashes on a computer, using Livermore Software Technology Corp.’s high-speed event simulator LS-DYNA. The software is sold through JSOL Corp., a Tokyo IT consultant and computer-aided engineering software distributor, and ESI Japan, a virtual prototyping software specialist. Toyota said in a statement earlier this month that the software is sold to other auto and parts manufacturers, but a spokesman declined to disclose which other automakers use it.
Effects on the simulated human bodies in the software are studied to improve devices such as airbags and improve safety performance functions, like seat shapes that reduce the risk of rib fractures in racing crashes, which NASCAR uses.
The three new child models are part of the Version 4 virtual set of virtual crash test dummies. The passenger and pedestrian child models are age 10 at 4 feet 6 inches tall , age 6 at 3 feet 10.5 inches and age 3 at 3 feet 1 inch .
The Toyota spokesman said the software’s model height and weight use the average U.S. size for those ages, based on data from the Centers for Disease Control and Prevention. There is no size consideration for other countries.
While physical crash tests have been using child-sized dummies for some time, the spokesman said virtual tests will be far more convenient and informational.
“It can take several days to prepare for and execute a single physical test,” the spokesman told Automotive News in an email. “In that same timeframe, however, dozens of virtual tests can be conducted.”
He added that the amount of data collected from a physical test is limited to the channels from data-collecting instruments on the dummies. Virtual models are not nearly as restricted in data collection.
Virtual vs. real
Virtual crash testing will probably never replace real-world crash tests.
The software can attempt to be perfect, but vehicles are not, said Russ Rader, a spokesman for the Insurance Institute for Highway Safety. The institute studies crashes to provide injury and vehicle-damage data for U.S. insurance and safety research and does not use virtual software.
Rader said once a vehicle is built, there can be differences between pre-production models and the finished product rolled off the assembly line. Only real human testing can account for those differences, he said.
“We do crash tests of actual vehicles because we’re trying to demonstrate for consumers the differences in how vehicles would protect them and their family in a real-world crash,” Rader told Automotive News. “We need to conduct tests of real vehicles to do that.”
The Toyota spokesman noted that both tests are necessary to find a correlation, but regulatory requirements rely heavily on the physical one.
Each previous version of Toyota’s software has added more features, such as face and bone structure, detailed brain simulation and internal organ interactions with the body during a crash. The last update was launched in 2015.
The updated software with the child models goes on sale this fall, Toyota said.
Researchers from Wayne State University in Detroit and the University of Michigan in Ann Arbor, Mich., collaborated with Toyota’s Technical Center to develop the new child-spec models, Toyota said.