Simulating human movement to optimize physical performance
Trained as mechanical engineer and neuroscientist, with a background in robotics, Scott Delp has long been fascinated with human movement.
As a professor of mechanical engineering and director of the Neuromuscular Biomechanics Lab at Stanford, Delp is the creator of OpenSim, an open source software application that simulates the complex choreography of nerve, bone and muscle that enables us to move.
OpenSim is designed to simulate how various segments of the body move relative to one another. It is available for download free-of-charge and includes extensive tutorials to help people learn how to use it. “You can start from scratch and be up and running in minutes,” says Delp, who has used OpenSim to study injuries, disease and athletic performance.
Delp first conceived of simulating human movement when he was a patient seeking help for a sports injury. He thought computer simulations would help doctors and engineers understand imperfect movement in order to cure or compensate for injuries and disabilities. Modeling would allow engineers to simulate injuries without risk to subjects and run thousands of virtual experiments to test alternatives.
OpenSim’s versatility has been a pleasant surprise. “One of the wonderful things about open source software is all the ways people find to use it that we either never thought of or don’t have the time to pursue,” says Delp, who is also a member of the Stanford Neurosciences Institute.
Many researchers have used OpenSim to study specific sections of the body or predict the outcomes of surgical interventions. One team is using the software to model human birth to understand why more defects occur in first births than in subsequent ones. Because many therapies are tested on lab animals as a prelude to human trials, other researchers have adapted OpenSim to model the locomotion of mice, rats and monkeys. One has even customized the tool to study Tyrannosaurus rex.
True to Delp’s original vision, his lab largely uses OpenSim to improve human movement and prevent injury. One technique his lab is working on uses computer-generated vibrations to provide sensory cues that help people to improve their walking or running gaits, like a little nudge here or there that help people improve movement.
Looking for new ways to get the big picture on human motion, Delp’s lab has begun to analyze the vast quantities of health data being amassed by smartphones and wearable devices. In one recent study published in the journal Nature, Delp and fellow Stanford professor Jure Leskovic, a computer scientist, used mobile data to analyze the physical activity of 717,000 men and women in 111 countries across the world.
They found that in areas where there is less physical activity – as measured in average steps per day – women tend to walk less than men and, consequently, suffer greater levels of obesity. “Like income inequality, it appears that activity inequality has a gender bias,” Delp says. The findings, he hopes, will inspire new public health campaigns or affect public policy to make cities more walkable, which reduces activity inequality.
For those interested in pursuing biomechanics as a career, Delp counsels academic diversity. While he is a mechanical engineer, he says the highly skilled people in his lab are cut from a wide variety of disciplines. His lab is always in need of people trained in biology, mathematics, neuroscience, computer science, design, and many other disciplines.
“It seems cross-training is good for both physical and scientific fitness,” Delp muses. “It takes all kinds to do this work.”