H∞boys and INDIans: Can they work together for flight control?
Event details
| Date | 07.10.2026 |
| Hour | 11:00 › 12:00 |
| Speaker | Dr. Spilios Theodoulis, associate professor in the Aerospace GNC cluster within the Control & Simulation (C&S) section, at the Faculty of Aerospace Engineering (AE) of the Delft University of Technology (TUDelft), Netherlands |
| Location | |
| Category | Conferences - Seminars |
| Event Language | English |
Abstract
Since the early days of aircraft autopilots, flight control law design has been dominated by gain scheduling: a family of linear controllers designed at trim points across the flight envelope and interpolated as the aerospace vehicle's operating condition changes. The approach is reliable and well understood, but it is also labor-intensive, requiring extensive linearization, controller interpolation, and lengthy re-tuning whenever the airframe or mission profile changes while its classical stability and performance guarantees hold only under restrictive assumptions.
(Incremental) Nonlinear Dynamic Inversion — (I)NDI — emerged as an attractive alternative, first as full-state NDI in the 1980s–90s and later, in its incremental form, replacing much (though not all) of the reliance on an accurate onboard model with direct sensor feedback of angular acceleration. The result is a simple, largely envelope-independent control structure that (almost) eliminates the scheduling and interpolation burden. The price, however, is the loss of the a priori robustness guarantees that gain-scheduled designs, however cumbersome, were built to provide. Robust control theory, born in the same timeframe as NDI, offers exactly those guarantees, but has traditionally lived in a separate, frequency-domain world from INDI's incremental, time-domain philosophy.
This talk shows how the two could be brought closer or even work together. We present hybrid INDI architectures combining a sensor-based and a model-based inversion core and paired with an outer loop tuned via structured synthesis therefore combining the advantages of both worlds. We then generalize this bridge through a formal duality between (I)NDI and quasi-LPV control, unifying robust analysis and synthesis across both worlds and clarifying exactly where INDI's modularity is paid for in achievable robustness. These results are validated across various aerospace applications such as aircraft, space launchers and drones.
Biography
Dr. Spilios Theodoulis is an associate professor in the Aerospace GNC cluster within the Control & Simulation (C&S) section, at the Faculty of Aerospace Engineering (AE) of the Delft University of Technology (TU Delft). Before joining TU Delft, he spent fourteen years at the French-German Research Institute of Saint-Louis (ISL), where he was Deputy Head of the GNC department. He is the founder of the Aerospace dynamics and RObust Control (AEROCON) research group along with its advanced projects branch—the ∞-Lab, and also an adjunct professor at the University of Strasbourg, University of Paris-Saclay and Cranfield University teaching tensor-based flight dynamics and automatic (flight) control. He has been involved in the research and development of both civilian and defense projects with EU government agencies, industry and academia in the field of GNC for more than 20 years.
His research interests focus on two complementary fields. First, modeling and flight dynamics of complex aerospace systems, including uncertainty modeling and linear/nonlinear-parameter-varying dynamics. Second, (multivariable) stability and control systems spanning from robust/nonlinear control and analysis to control theory-inspired guidance algorithms. The algorithms developed are showcased in several classes of systems (including the facilities of the AE C&S section such as SIMONA, PH-LAB, etc.) such as civil and fighter aircraft, rotorcraft, drones, hypersonic vehicles, space launchers, as well as guided systems. He maintains active collaborations with leading research institutes, including ONERA, DLR, NLR, TNO, NASA, and ESA-ESTEC; academic partners such as ISAE-SUPAERO, Cranfield University, and NYU Abu Dhabi; and industrial partners including Dassault Aviation, Airbus Defence & Space, MBDA, Indra Deimos, PLD Space, Lockheed Martin Skunk Works, and others.
He is an Associate Fellow of the American Institute of Aeronautics and Astronautics (Class of 2023), co-recipient of both 1-Star and 2-Star Innovation Awards (2019) from MBDA, and his students have received (some) distinctions, including the best PhD award in the area of systems and control from the University of Strasbourg (2017). He also serves on various technical committees (AIAA GNC, IFAC Aerospace, CEAS GNC) as well as on the organizing and program conference committees related to the field of GNC (AIAA SciTech, IFAC WC/ACA, EuroGNC, etc.).
His most daunting challenge to date though, remains the robust stabilization of his two young daughters Iris and Hera, who are constantly reminding him that nature is neither linear nor time-invariant.
Since the early days of aircraft autopilots, flight control law design has been dominated by gain scheduling: a family of linear controllers designed at trim points across the flight envelope and interpolated as the aerospace vehicle's operating condition changes. The approach is reliable and well understood, but it is also labor-intensive, requiring extensive linearization, controller interpolation, and lengthy re-tuning whenever the airframe or mission profile changes while its classical stability and performance guarantees hold only under restrictive assumptions.
(Incremental) Nonlinear Dynamic Inversion — (I)NDI — emerged as an attractive alternative, first as full-state NDI in the 1980s–90s and later, in its incremental form, replacing much (though not all) of the reliance on an accurate onboard model with direct sensor feedback of angular acceleration. The result is a simple, largely envelope-independent control structure that (almost) eliminates the scheduling and interpolation burden. The price, however, is the loss of the a priori robustness guarantees that gain-scheduled designs, however cumbersome, were built to provide. Robust control theory, born in the same timeframe as NDI, offers exactly those guarantees, but has traditionally lived in a separate, frequency-domain world from INDI's incremental, time-domain philosophy.
This talk shows how the two could be brought closer or even work together. We present hybrid INDI architectures combining a sensor-based and a model-based inversion core and paired with an outer loop tuned via structured synthesis therefore combining the advantages of both worlds. We then generalize this bridge through a formal duality between (I)NDI and quasi-LPV control, unifying robust analysis and synthesis across both worlds and clarifying exactly where INDI's modularity is paid for in achievable robustness. These results are validated across various aerospace applications such as aircraft, space launchers and drones.
Biography
Dr. Spilios Theodoulis is an associate professor in the Aerospace GNC cluster within the Control & Simulation (C&S) section, at the Faculty of Aerospace Engineering (AE) of the Delft University of Technology (TU Delft). Before joining TU Delft, he spent fourteen years at the French-German Research Institute of Saint-Louis (ISL), where he was Deputy Head of the GNC department. He is the founder of the Aerospace dynamics and RObust Control (AEROCON) research group along with its advanced projects branch—the ∞-Lab, and also an adjunct professor at the University of Strasbourg, University of Paris-Saclay and Cranfield University teaching tensor-based flight dynamics and automatic (flight) control. He has been involved in the research and development of both civilian and defense projects with EU government agencies, industry and academia in the field of GNC for more than 20 years.
His research interests focus on two complementary fields. First, modeling and flight dynamics of complex aerospace systems, including uncertainty modeling and linear/nonlinear-parameter-varying dynamics. Second, (multivariable) stability and control systems spanning from robust/nonlinear control and analysis to control theory-inspired guidance algorithms. The algorithms developed are showcased in several classes of systems (including the facilities of the AE C&S section such as SIMONA, PH-LAB, etc.) such as civil and fighter aircraft, rotorcraft, drones, hypersonic vehicles, space launchers, as well as guided systems. He maintains active collaborations with leading research institutes, including ONERA, DLR, NLR, TNO, NASA, and ESA-ESTEC; academic partners such as ISAE-SUPAERO, Cranfield University, and NYU Abu Dhabi; and industrial partners including Dassault Aviation, Airbus Defence & Space, MBDA, Indra Deimos, PLD Space, Lockheed Martin Skunk Works, and others.
He is an Associate Fellow of the American Institute of Aeronautics and Astronautics (Class of 2023), co-recipient of both 1-Star and 2-Star Innovation Awards (2019) from MBDA, and his students have received (some) distinctions, including the best PhD award in the area of systems and control from the University of Strasbourg (2017). He also serves on various technical committees (AIAA GNC, IFAC Aerospace, CEAS GNC) as well as on the organizing and program conference committees related to the field of GNC (AIAA SciTech, IFAC WC/ACA, EuroGNC, etc.).
His most daunting challenge to date though, remains the robust stabilization of his two young daughters Iris and Hera, who are constantly reminding him that nature is neither linear nor time-invariant.
Practical information
- General public
- Free
Organizer
- Professor Alireza Karimi