Speaker: Jack M Wang
Title: Simulating Human Locomotion: Optimization, Uncertainty, and Biomechanics
Locomotion, specifically walking and running, are common and essential human movements. The ability to create physically and biomechanically plausible simulations of locomotion is of interest to applications ranging from game content creation to pathological gait analysis, and can contribute to our understanding of motor control.
However, while humans can move on varied terrains, start and stop on a dime, and recover from trips with ease, getting simulated humanoids to simply walk forward without falling is a challenging task. Achieving locomotion requires solving a high-dimensional, nonlinear, and underactuated control problem. Furthermore, out of all the control strategies that accomplish the task, how do we select one that produces human-like movements? In this talk, the speaker will present an approach to simulate and control 3D humanoid locomotion that produces results matching human data to a much greater extent than previous state-of-the-art.
First, the speaker would describe a method for synthesizing 3D walking controllers through optimization, and show that given a well-designed set of objectives and control parameterization, walking controllers can be synthesized without the need for a stable initialization nor motion capture data. Second, the speaker would explore the effect of environmental uncertainty on control optimization, and show that more robust controllers with cautious-looking styles automatically emerge when factors such as external pushing forces and random user inputs are modeled in the simulation. Finally, the speaker would demonstrate the importance of biomechanics in recovering human-like strategies for locomotion. Specifically, the use of muscle models to ensure biologically plausible forces measurably increases the realism of motions generated by optimized controllers.