A&A demonstrated its growing prominence in aerospace control systems, with two doctoral students capturing top honors at the 2024 Conference on Decision and Control (CDC) in Milan, one of the premier venues in control and systems theory.

Spencer Kraisler and Mohammad Al-Jarrah each received Outstanding Paper Awards, with Kraisler additionally winning the Best Student Paper Award – a noteworthy achievement for the department at one of the most prestigious conferences in the field. The CDC student awards are highly coveted and competitive with doctoral students nominated from around the world and a rigorous selection process.

"These awards reflect the innovative work happening in our controls research groups," said Professor Mehran Mesbahi, who directs the Robotics, Aerospace Information, and Network Systems (RAIN) Laboratory where Kraisler conducts his research. "Spencer's work on geometric approaches to control synthesis combines some of the best foundational traditions in control theory with modern approaches to learning and optimization. His significant contribution to the field has been to tackle the intricate geometrical aspects of dynamic feedback design in the context of policy optimization and how to use this geometry for developing novel efficient algorithms.”

Working with Assistant Professor Amir Taghvaei, Al-Jarrah and his co-authors, including UW applied math Assistant Professor Bamdad Hosseini, developed advanced methods for estimating the state of aerospace systems using real-world sensor data. Their approach improves the scalability and accuracy of estimation algorithms impacting applications ranging from aircraft navigation to space debris tracking.

"Mohammad's work shows how modern computational techniques from machine learning and data science could transform the field of estimation and nonlinear filtering" noted Taghvaei. "His method makes it possible, for the first time, to implement a nonlinear filter in a model-free and data-driven setup, improving scalability and computational efficiently while maintaining high accuracy."

The strong showing at CDC also highlighted the department's broader impact in the controls community, with several of our faculty and current students presenting their research at the conference. As the aerospace industry continues its rapid evolution, UW's aerospace program remains at the forefront of developing the theoretical foundations and practical tools that will shape the future of flight.

Go to the sources

Data-Driven Approximation of Stationary Nonlinear Filters with Optimal Transport Maps, by Mohammad Al-Jarrah, Bamdad Hosseini and Amir Taghvaei.

Abstract: The nonlinear filtering problem is concerned with finding the conditional probability distribution (posterior) of the state of a stochastic dynamical system, given a history of partial and noisy observations. This paper presents a data-driven nonlinear filtering algorithm for the case when the state and observation processes are stationary. The posterior is approximated as the push-forward of an optimal transport (OT) map from a given distribution, that is easy to sample from, to the posterior conditioned on a truncated observation window. The OT map is obtained as the solution to a stochastic optimization problem that is solved offline using recorded trajectory data from the state and observations. An error analysis of the algorithm is presented under the stationarity and filter stability assumptions, which decomposes the error into two parts related to the truncation window during training and the error due to the optimization procedure. The performance of the proposed method, referred to as optimal transport data-driven filter (OT-DDF), is evaluated for several numerical examples, highlighting its significant computational efficiency during the online stage while maintaining the flexibility and accuracy of OT methods in nonlinear filtering.

Output-Feedback Synthesis Orbit Geometry: Quotient Manifolds and LQG Direct Policy Optimization, by Spencer Kraisler and Mehran Mesbahi.

Abstract: We consider direct policy optimization for the linear-quadratic Gaussian (LQG) setting. Over the past few years, it has been recognized that the landscape of dynamic output-feedback controllers of relevance to LQG has an intricate geometry, particularly pertaining to the existence of degenerate stationary points, that hinders gradient methods. In order to address these challenges, in this paper, we adopt a system-theoretic coordinate-invariant Riemannian metric for the space of dynamic output-feedback controllers and develop a Riemannian gradient descent for direct LQG policy optimization. We then proceed to prove that the orbit space of such controllers, modulo the coordinate transformation, admits a Riemannian quotient manifold structure. This geometric structure–that is of independent interest–provides an effective approach to derive direct policy optimization algorithms for LQG with a local linear rate convergence guarantee. Subsequently, we show that the proposed approach exhibits significantly faster and more robust numerical performance as compared with ordinary gradient descent.