On September 6, 2024, Brian Plancher, assistant professor of computer science, received a grant, totaling over $590,000, from the National Science Foundation (NSF) to support a new project developing open-source, GPU-accelerated optimization cyberinfrastructure targeting edge robotics applications. This cyberinfrastructure will help researchers leveraging optimization to overcome fundamental shortcomings in current software, with a particular focus in optimal control applications for robotics.
These optimal control solvers, Plancher explains, have a diverse range of applications, from robotics to manufacturing and utilities, continuously optimizing robot motions, plant operations, and power grid networks. Unfortunately, current software infrastructure fundamentally limits the performance of many of these systems due to its inability to effectively scale to large-scale problems. Accelerating this software on GPUs will enable robots to make better decisions faster, enabling safer and more dynamic real-world operation. Through the development of this cyberinfrastructure, Plancher intends to close this gap by addressing the critical scientific need for low-latency optimal control at the edge, as well as developing unified APIs and benchmark problems and datasets for easy adoption and generalization by the broader research community.
Using the over half a million dollars of funding support from the NSF, Plancher will spend the next three years developing more general, accessible, and documented open-source GPU optimal control libraries that can support a broader range of scientific research, with a focus on robotics. His research will culminate with the creation of an integrated educational curriculum that will enable researchers and practitioners worldwide to learn about these topics and continue developing this exciting interdisciplinary area of computer science and engineering.