Parasol Lab at the University of Illinois conducts research at the intersection of robotics, algorithms, and parallel computing, with an emphasis on the foundations that enable computation at scale. Our work is unified by a core theme: developing structured representations and principled methods that make complex, high-dimensional problems more tractable, both in robotic decision-making and in parallel execution on modern architectures.

In robotics, we study motion planning, multi-robot coordination, and task and motion planning, with particular attention to environments where combinatorial complexity and geometric constraints interact. A recurring strategy is to introduce structure: through graph and hypergraph formulations, decomposition into reusable subproblems, and experience- or database-guided planning. These structural impositions help us to better characterize the search space and to obtain algorithms with improved performance and reliability. We also investigate human-centered approaches, including XR-based demonstration and interaction, as a way to formalize intent and connect human input to algorithmic planning and learning.

In parallel computing, we develop compiler and runtime techniques that expose and manage concurrency for irregular workloads, aiming to improve programmability while delivering high performance. This includes research on task-parallel execution for modern computational pipelines, such as dataflow graphs arising in machine learning, as well as high-performance methods for scientific computing and data-intensive applications. Across these efforts, we focus on rigorous algorithmic analysis, careful empirical evaluation, and open questions in how to make parallelism more automatic, predictable, and portable.

Overall, Parasol Lab’s research advances the theory and practice of scalable computation, contributing new algorithmic frameworks, system techniques, and empirical insights that support robotics, parallel programming, and scientific computing.

Join us!