Dr.Puneet Singla
Dr. Puneet Singla is the Harry and Arlene Schell Professor of Aerospace Engineering at The Pennsylvania State University, where he directs the Control, Autonomy and Space Systems Laboratory. He received his B.Tech. from IIT Kanpur and his Ph.D. from Texas A&M University, both in Aerospace Engineering.
His research focuses on developing autonomous decision-making algorithms for operations in contested and highly uncertain environments. The central theme of his work is enabling engineered systems to sense, reason, decide, and adapt in real-time despite incomplete, uncertain, or deceptive information. By combining physics-based modeling with uncertainty quantification, estimation, and control theory, his research develops computationally tractable methods that connect prediction directly to action. These ideas are applied to space domain awareness, cislunar and multi-body operations, autonomous spacecraft, rendezvous and proximity operations, in-space servicing and assembly, and hypersonic systems.
Dr. Singla is a Fellow of AIAA and AAS and a recipient of the NSF CAREER Award, AFOSR Young Investigator Award, and IEEE AESS Judith A. Resnik Space Award.
Uncertainty is intrinsic to nearly every engineering dynamical system. Models are imperfect, measurements are noisy and incomplete, and disturbances are rarely known precisely. Yet useful decisions must often be made in real time by combining evolving models with heterogeneous and imperfect data. A central challenge is therefore to develop computational frameworks that can characterize uncertainty accurately, assimilate information dynamically, and translate that information into sensing and control decisions without becoming computationally prohibitive.
This talk will present an overview of research efforts in the Control, Autonomy and Space Systems (CASS) Laboratory at Penn State aimed at developing computationally tractable methods for nonlinear dynamical systems operating under uncertainty. A central theme is the construction of deterministic numerical representations of probability distributions that retain the essential statistical information required for prediction and decision making while avoiding prohibitively large Monte Carlo ensembles. The talk will introduce the Conjugate Unscented Transformation (CUT), a family of non-product quadrature rules designed to efficiently capture higher-order moments, and will show how these ideas can be combined with sparse approximation methods for uncertainty propagation through nonlinear systems.
These computational ideas naturally extend to dynamic model-data interaction. The talk will describe methods for solving the Fokker-Planck-Kolmogorov equation for the evolution of probability density functions, Bayesian approaches for integrating heterogeneous sensor measurements with dynamical models, and computational methods for solving Hamilton-Jacobi-Bellman equations for optimal feedback control. Together, these tools provide a common mathematical foundation for connecting modeling, uncertainty quantification, estimation, dynamic sensing, planning, and control.
A recurring theme will be the dynamic coupling between models, data, and decisions, which lies at the heart of the DDDAS paradigm. Measurements are not simply used to correct a model after the fact. Instead, evolving uncertainty and model predictions can determine where, when, and what to measure next, while newly acquired data continually refine the state, model, and subsequent decisions. This feedback between computation, sensing, and action enables systems that can adapt their information-gathering and control strategies as conditions evolve.
The talk will highlight applications ranging from resident space object tracking and space domain awareness to uncertainty-aware navigation, trajectory design, and autonomous spacecraft operations. These examples illustrate how explicitly accounting for uncertainty can lead not only to improved prediction, but also to more informative sensing, more robust decision making, and ultimately more capable autonomous systems.
