Yu Liu

Bio: Yu Liu is an academic researcher from University of Virginia. The author has contributed to research in topics: Adaptive control & Control system. The author has an hindex of 11, co-authored 25 publications receiving 343 citations. Previous affiliations of Yu Liu include National Institute of Aerospace & Ford Motor Company.

Modeling and Model Reference Adaptive Control of Aircraft with Asymmetric Damage

Abstract: This paper addresses some fundamental issues in adaptive control of aircraft with struc- tural damage It presents a thorough study of linearized aircraft models with damage to obtain new details of system descriptions, such as coupling and partial derivatives of lateral and longitudinal dynamics A detailed study of system invariance under damage conditions is performed for generic aircraft models to obtain key system characterizations for model reference adaptive control (MRAC), such as infinite zero structure and signs of high frequency gain matrices A comprehensive study of multivariable MRAC systems in the presence of damage is performed to obtain critical design specifications for adaptive flight control, such as system and controller parametrizations and adaptive parameter up- date laws Both analytical and simulation results are given to illustrate the design and performance of adaptive control systems for aircraft flight control Adaptive control of aircraft in the presence of damage has been an important topic in the research of flight control design for aircraft safety Damage can cause uncertain parametric and structural variations, which requires new aircraft modeling and control approaches In Reference (1), a study of aircraft dynamics with damage is presented, and a neural network based adaptive control algorithm is introduced for control of aircraft in the presence of structure uncertainties In (2), equations of motion are introduced in detail for aircraft with asymmetric mass loss In (3), we introduced a nonlinear aircraft model with partial wing damage, and illustrated linearization of such a model In (4), real time identification of a damaged aircraft model is studied A two-step identification process is introduced, which consists of an aircraft state estimation phase and an aerodynamic model identification step With such a two-step process, the nonlinear part of the model identification is isolated in the first phase, and the aerodynamic parameter identification procedure is simplified to a linear one A hybrid adaptive control method is proposed in (5) for control of aircraft with damage The control design is based on a neural network parameter estimation blended with a direct adaptive law A stability and convergence analysis is presented for this adaptive control methodology For accommodating unknown changes in the structure and parameters, multivariable MRAC designs offer many advantages In (6), we introduced an MRAC design based on the LDS decomposition of the high frequency gain matrix for the control of aircraft with multiple wing damage The key design conditions are that, both the nominal and post-damage systems should have a uniform known modified interactor matrix, and the leading principal minors of their high frequency gain matrices should be nonzero with their signs unchanged In (7), we studied linearization of nonlinear aircraft models under damage conditions and designed a multivariable MRAC scheme which does not require the knowledge of the signs of the high frequency gain matrix Potential extension to aircraft flight control systems with changing signs of the high frequency gain matrix remains a topic of future research

Multivariable Adaptive Control of NASA Generic Transport Aircraft Model with Damage

Abstract: DOI: 10.2514/1.53258 This paper develops a linearization-based adaptive control technique for control of the nonlinear NASA Generic Transport Model with large parametric and structural uncertainties caused by damage. It presents new studies of linearization of the nonlinear aircraft dynamics in the presence of damage. A state-feedback multivariable model reference adaptive control scheme is developed for the linearized aircraft modelwith damageto ensure stability and asymptotic output tracking for aircraft in the presence of damage. The invariance characteristic of system infinityzerostructuresofthenonlinearandlinearsystemmodelsisinvestigated.Thisworkshowshowalinearization-based modelreference adaptivecontrol schemeusingstate feedbackcan beapplied to anonlinearaircraft dynamicsystem with damage, and the control system simulation of the Generic Transport Model demonstrates some desired performance of such a control system in the neighborhood of the chosen operating point.

Multivariable MRAC with state feedback for output tracking

Abstract: This paper revisits the multivariable MRAC problem, by studying adaptive state feedback control for output tracking of multi-input and multi-output (MIMO) systems. With such a control scheme, the plant-model matching condition is much less restrictive than those for state tracking, while the controller enjoys a simpler structure than that of an output feedback design with the guarantee of the asymptotic tracking of multiple outputs. Such a control scheme is useful for applications when the plant-model matching condition for state tracking cannot be satisfied. A stable adaptive control scheme is developed based on LDS decomposition of high frequency gain matrix, which ensures closed-loop stability and asymptotic output tracking. A simulation study is conducted for an aircraft model, with desired simulation results presented.

Adaptive Compensation of Aircraft Actuation Failures Using an Engine Differential Model

Abstract: This paper investigates actuator failure compensation for aircraft flight control in a novel framework. A general failure compensation scheme for asymptotic tracking is developed based on a direct adaptive control approach. This control scheme is capable of utilizing the remaining control authority to achieve the desired performance in the presence of unknown and uncertain constant actuator failures occurring at unknown time instants. A nonlinear aircraft model that incorporates independently adjustable engine throttles and ailerons is employed and linearized to describe the aircraft's longitudinal and lateral motion. This model captures the key features of aircraft flight dynamics when in the engine differential mode. The proposed control scheme is applied to a transport aircraft model in the presence of three types of failures during operation: rudder failure, aileron failure, and engine malfunction. Simulation results are presented to assess the effectiveness of this adaptive failure compensation design.

Multivariable MRAC using Nussbaum gains for aircraft with abrupt damages

Abstract: This paper derives a stable multivariable model reference adaptive control (MRAC) scheme for systems with abrupt parameter variations (which may cause uncertain sign changes in the system?s high frequency gain matrix), motivated by the application to in-flight aircraft systems with damages. Such sign changes are illustrated by an aircraft model with asymmetric abrupt damages, and their uncertainty is handled by a Nussbaum gain based adaptive control design to control the aircraft for both healthy and post-damage situations, by adapting controller parameters autonomously after the damages occur, for which the knowledge of time instants, structures and values of the damages is not required. A piecewise continuous Lyapunov function is utilized to prove the desired system stability and tracking properties in the presence of damages.

Fault detection in finite frequency domain for Takagi-Sugeno fuzzy systems with sensor faults.

Abstract: This paper is concerned with the fault detection (FD) problem in finite frequency domain for continuous-time Takagi-Sugeno fuzzy systems with sensor faults. Some finite-frequency performance indices are initially introduced to measure the fault/reference input sensitivity and disturbance robustness. Based on these performance indices, an effective FD scheme is then presented such that the generated residual is designed to be sensitive to both fault and reference input for faulty cases, while robust against the reference input for fault-free case. As the additional reference input sensitivity for faulty cases is considered, it is shown that the proposed method improves the existing FD techniques and achieves a better FD performance. The theory is supported by simulation results related to the detection of sensor faults in a tunnel-diode circuit.

Flight Dynamics and Hybrid Adaptive Control of Damaged Aircraft

Abstract: This paper presents a recent study to investigate flight dynamics and adaptive control methods for stability and control recovery of a damaged generic transport aircraft. Aerodynamic modeling of damage effects is performed using an aerodynamic code to assess changes in the stability and control derivatives of the damaged aircraft. Flight dynamics for a general aircraft are developed to account for changes in aerodynamics, mass properties, and the center of gravity that can compromise the stability of the damaged aircraft An iterative trim analysis is developed to compute incremental trim states. A neural network hybrid direct-indirect adaptive flight control is developed for the stability augmentation control of the damaged aircraft. The proposed method performs an online estimation of damaged plant dynamics to improve the command tracking performance in conjunction with a direct adaptive controller. The plant estimation is based on two approaches: 1) an indirect adaptive law derived from the Lyapunov stability theory to ensure that the tracking error is bounded, and 2) a recursive least-squares method that minimizes the modeling error. Simulations show that the hybrid adaptive controller can provide a significant improvement in the tracking performance over a direct adaptive controller working alone.

Adaptive disturbance rejection using ARMARKOV/Toeplitz models

Abstract: An adaptive disturbance rejection algorithm is developed for the standard control problem The multiple input-multiple output (MIMO) system and controller are represented as ARMARKOV/Toeplitz models, and the parameter matrix of the compensator is updated online by means of a gradient algorithm The algorithm requires minimal knowledge of the plant, specifically, the numerator of the ARMARKOV model of the transfer function from the control inputs to the performance variables is required No knowledge about the spectrum of the disturbance is needed Experimental results demonstrating tonal and broadband disturbance rejection in an acoustic duct are presented