By Tom Avril
The Philadelphia Inquirer
A blur of motion on the ice, Haley Beavers spun into the air to attempt a triple loop - one of the standard jumps in the repertoire of an elite skater.
She did not quite make it, stumbling on her landing at the University of Delaware.
No matter. Within minutes, a sleek computer simulation showed the 14-year-old that if she drew her arms in just a bit closer, she would be golden.
Call it better skating through physics. Four of the five U.S. Olympic singles skaters competing this month in Russia have used the simulation software, in a joint project between researchers at Delaware and a Maryland company called C-Motion Inc. The next generation of Olympic hopefuls, such as Beavers, has lined up as well, using it to practice more efficiently and, ideally, cut down on the number of bone-jarring falls and injuries.
“I thought it was amazing,” said U.S. team member Jason Brown, who first used the program when he was learning the triple axel two years ago. “It scientifically showed I could be able to do the jump.“
Coaches have been filming athletes’ performances almost since the invention of the video camera, playing back footage in slow motion so they can suggest a better position for an arm or leg.
With the simulation software, you can actually arrange the position of the athlete’s limbs on the computer screen, then play it back to see what would happen in real life.
The process starts with motion-capture technology, similar to what is used in Hollywood special-effects and animation studios.
Delaware biomechanics professor Jim Richards, the project director, also is using this technology to improve treatment of children with a rare shoulder injury - a joint effort with Shriners Hospital for Children in Philadelphia.
One day last month, two of Richards’ Ph.D. students, Garry Quinton and Tyler Richardson, attached 33 silvery, cube-shaped markers all over Haley Beavers’ bodysuit, head, and skates.
A suite of 10 cameras hung from the ceiling of the university’s Gold Ice Arena. Next to each camera lens was a bright red light, aimed down at the ice.
The silvery markers on the skater’s body reflected the light back to the cameras. That way the cameras and the software connected to them could record exactly where each of the 33 dots was located, and thus the location of each limb.
“We can literally re-create the position of that limb at any point in time,” Richards said.
The cameras recorded Beavers as she tried a few jumps. Then she stepped off the ice into a small room to one side of the rink, and the heavy data-crunching began.
Ph.D. student Kristen Nicholson sat at the computer keyboard in the frosty air, wearing gloves with the fingertips cut off to enable easier typing.
Nicholson asked the skater for her height and weight - a shade under 5-foot-1 and 103 pounds - and entered it into the computer so the program could simulate the physics of her dizzying spins.
Within moments, two mannequin-like avatars, or figures, appeared on a nearby screen. One represented Haley Beavers as she really was, captured at a rate of 250 frames per second. The other mannequin was a mathematical model, on which Richards could move the arms and legs to see if it would make her spin faster.
Richards played back the animation of Beavers’ actual triple loop, and saw that she was about one-fifth of a turn short of the necessary three revolutions.
The teenager and her coach, four-time national pairs skating champion Ron Ludington, agreed that her elbows were sticking out too far.
“It’s loose,” Beavers said of her form.
“Definitely get those arms in,” Ludington said.
We’ve all seen how skaters spin faster when they hold their arms in tighter. But predicting the exact speed for each fraction of a second requires some fancy footwork.
The skater’s rotational velocity depends on the distribution of mass in relation to the center of her spin. The more mass that is located away from the center, say, with an outstretched arm, the slower her rotation.
The C-Motion software, written by company scientific director and former Delaware researcher Tom Kepple, relies on estimates for the mass of each body part. Taking into account the location of each segment, it then calculates the skater’s speed for each 250th of a second.
Richards used a mouse to pull in the arms on the model skater by a few degrees, then played it back. Presto! The virtual Haley Beavers added an extra quarter-turn to her jump.
“You should be able to do this,” Richards told the skater. “It’s a matter of staying on target.”
The motion-capture technology is an off-the-shelf product made by another company. Kepple’s skating simulator is a separate, custom piece of software.
The project started five years ago, under contract with U.S. Figure Skating. The software has been in usable form since 2011, though Kepple said he had continued to tweak it after that.
Kepple and Richards said they did not know if other nations’ figure-skating teams were using anything like it. In England, Loughborough University researchers have used a similar approach to advise gymnasts, divers, trampolinists, and freestyle aerial skiers.
Ludington, a member of the figure skating World Hall of Fame, is a convert.
“It’s invaluable,” Ludington said. “We’ve all changed, the coaches who are teaching at this level, because they see more. It just improves the sport. That’s why skaters are getting better and better and better, and at a younger age.”
The research at Shriners Hospital, meanwhile, is still in development. Richards’ lab is working with Shriners physicians Dan Zlotolow and Scott Kozin, using motion-capture technology to study the movements of children with a shoulder ailment called brachial plexus birth palsy.
Zlotolow said the research already was helping him determine the best course for surgery on such children, whose neck nerves have become stretched during birth.
“We want to know where to put these muscles to best move the shoulder,” Zlotolow said.
The skating simulation has now been used by 80 skaters. Kepple said it enabled skaters to master their jumps with fewer attempts, and thus ideally would lead to less injury.
“It makes their practice more focused,” Kepple said. “I like to describe it as, they kind of leave here with a prescription for their training.”
But it can take time to go from simulation to success. Brown, the 2014 Olympian, came back to Delaware last spring to work on a quadruple jump, and is not ready to include it in his routines this month in Russia.
On the computer screen, however, he has already nailed it.