Understanding the evolutionary pressures that shaped the human lineage requires an understanding of complex, dynamic interactions between multiple selection pressures, species' anatomy and locomotor control, and the environment. Evolutionary modeling using computer simulation is well-suited to this type of investigation because it is inherently dynamic and quantitative. However, it remains an underutilized approach in biological anthropology.
This project will harness recent advances in computer science and machine learning to develop and validate the techniques underpinning a new approach for investigating locomotor control and evolutionary change in chimpanzees, humans, and extinct hominins. Specifically, the project will develop computational optimization techniques for use in the open-source OpenSim modeling and simulation framework. These techniques, coupled with forward dynamics simulations of complex musculo-skeletal systems (such as humans), will be used to learn the best gait for a given optimization criterion, such as minimizing metabolic cost. The resulting gait, along with the force data generated by the simulation, will be compared to data obtained from the measurement of live subjects to ensure the validity of the approach.
This pilot project focusing on modeling human locomotion is a first step towards a fuller investigation of the control regimes and morphological optimization of chimpanzees, humans, and fossil hominins. Understanding the control regimes (e.g. minimizing cost, maximizing stability, or speed) underlying human and chimpanzee gait is critical for reconstructing locomotor evolution in hominins and hominoids.
