Jonathan A. Muse

A Loop Recovery Method for Adaptive Control

This paper presents a new modification term for use in adaptive control to improve an already existing design. By employing this term in a conventional adaptive law, the loop transfer properties of a reference model associated with a non-adaptive control design can be preserved. Consequently, this term increases the level of confidence of adaptive flight control systems for purposes of increased flight safety. The results are illustrated on an unmanned combat aerial vehicle dynamic model. I. Introduction n this paper we present an improved method of adaptation that enhances robustness of adaptive control systems. The method is arrived at by examining the loop transfer properties of an adaptive system when linearized about a given flight condition. The design approach modifies a conventional adaptive law with the goal of preserving the loop transfer properties of a reference model associated with a non-adaptive control design. The aim is to achieve an adaptive system that preserves the stability margins of a non-adaptive design, while at the same time providing the benefits of adaptation to modeling error. I

Adaptive Output Feedback Based on Closed-Loop Reference Models for Hypersonic Vehicles

This paper presents a new method of synthesizing an output feedback adaptive controller for a class of uncertain, nonsquare multi-input/multioutput systems that often occur in hypersonic vehicle models. The main challenge that needs to be addressed is the determination of a corresponding square and strictly positive real transfer function. This paper proposes a new procedure to synthesize two gain matrices that allow the realization of such a transfer function, thereby allowing a globally stable adaptive output feedback law to be generated. The unique features of this output feedback adaptive controller are a baseline controller that uses a Luenberger observer, a closed-loop reference model, manipulations of a bilinear matrix inequality, and the Kalman–Yakubovich lemma. Using these features, a simple design procedure is proposed for the adaptive controller, and the corresponding stability property is established. The proposed adaptive controller is compared to the classical multi-input/multioutput adaptiv...

Circumnavigation of an unknown target using UAVs with range and range rate measurements

This paper presents two control algorithms enabling a UAV to circumnavigate an unknown target using range and range rate (i.e., the derivative of range) measurements. Given a prescribed orbit radius, both control algorithms (i) tend to drive the UAV toward the tangent of prescribed orbit when the UAV is outside or on the orbit, and (ii) apply zero control input if the UAV is inside the desired orbit. The algorithms differ in that, the first algorithm is smooth and unsaturated while the second algorithm is non-smooth and saturated. By analyzing properties associated with the bearing angle of the UAV relative to the target and through proper design of Lyapunov functions, it is shown that both algorithms produce the desired orbit for an arbitrary initial state. Two examples are provided as a proof of concept.

Validation of A Low-Order Model for Closed Loop Flow Control Enabled Flight

A simple low-order model is derived to determine the flow forces and moments on an airfoil that arbitrarily pitches and plunges with the presence of synthetic jet actuation for the use in an adaptive closed-loop control scheme. The low-order model captures the attached flow response of an airfoil in the presence of synthetic jet actuators near the trailing edge. The model includes two explicit non-linear states for fluid variables and can be easily coupled to the rigid body dynamics of the system. The model is validated with high fidelity numerical simulations and experiments. The low-order model agreement with experiments is good for low reduced frequency pitching. The agreement to numerical simulations is also good for reduced frequencies that are an order of magnitude higher than those attainable in experiments.

Adaptive Control of a Generic Hypersonic Vehicle

This paper presents an adaptive augmented, gain-scheduled baseline LQR-PI controller applied to the Road Runner six-degree-of-freedom generic hypersonic vehicle model. Uncertainty in control effectiveness, longitudinal center of gravity location, and aerodynamic coefficients are introduced in the model, as well as sensor bias and noise, and input time delays. The performance of the baseline controller is compared to the same design augmented with one of two different model-reference adaptive controllers: a classical openloop reference model design, and modified closed-loop reference model design. Both adaptive controllers show improved command tracking and stability over the baseline controller when subject to these uncertainties. The closed-loop reference model controller offers the best performance, tolerating a reduced control effectiveness of 50%, rearward center of gravity shift of up to -1.6 feet (11% of vehicle length), aerodynamic coefficient uncertainty scaled 4× the nominal value, and sensor bias of up to +3.2 degrees on sideslip angle measurement. The closed-loop reference model adaptive controller maintains at least 70% of the delay margin provided by the robust baseline design when subject to varying levels of uncertainty, tolerating input time delays of between 15-41 ms during 3 degree angle of attack doublet, and 80 degree roll step commands.

A 1-DOF Wind Tunnel Experiment in Adaptive Flow Control

This paper presents an adaptive control approach for pitch control of a one degree-of- freedom wing using imbedded flow control actuators. Such actuators offer a unique opportunity for rapid maneuvering and gust rejection with low power consumption. The proposed adaptive non-model based control approach is motivated by the difficulty in constructing a reasonably accurate physical model of active flow control actuators. In addition, during dynamic maneuvers, possible interactions between unsteady fluid dynamics and vehicle dynamics introduce additional uncertainty. The control architecture employs a neural network based adaptive element that permits adaptation to both parametric uncertainty and unmodeled dynamics. Wind tunnel experimental results and controller implementation details are presented.

Nonlinear Adaptive Dynamic Inversion Applied to a Generic Hypersonic Vehicle

Flight control of hypersonic vehicles is challenging because of the wide range of operating conditions encountered and certain aspects unique to high speed flight. A particular safety concern in hypersonic flight is the risk of an inlet unstart, which not only produces a significant decrease in thrust but also results in a change to the aerodynamics and thus can lead to the loss of the vehicle. Previous work on control design for hypersonic vehicles often uses linearized or simplified nonlinear dynamical models of the vehicle, and very little work has been done on recovering from unstart events. Using a generic hypersonic vehicle as a control design and simulation model, this paper develops a nonlinear adaptive dynamic inversion control architecture with a control allocation scheme to track realistic flight path angle trajectories. A robustness analysis is performed on the initial control architecture design, which shows that the control architecture is able to handle time delays, perturbations in stability derivatives, and reduced control surface effectiveness. The control architecture then is evaluated for its ability to handle inlet unstart. Simulation results presented in the paper demonstrate that the approach achieves desired tracking performance while being robust to the particular uncertainties and inlet unstart conditions studied.

Adaptive Attitude and Vibration Control of the NASA Ares Crew Launch Vehicle

The control system of the NASA Crew Launch Vehicle is designed to meet performance and robustness requirements during its ascent ∞ight phase. However, the controller bandwidth and the attainable level of robust performance is limited by the degree of ∞exibility inherent in the long and slender design that has been adopted for this vehicle. Since there remains a substantial degree of uncertainty regarding the structural dynamics of this vehicle, the degree of risk associated with ∞ight control is reduced by permitting a greater level of robust performance to be attained by augmenting the existing ∞ight control system design with an adaptive element. Attitude stabilization and ∞exible mode suppression of the model is achieved by using a model reference adaptive controller designed to maintain nominal tracking performance in the presence of disturbances, parametric uncertainties and unmodeled dynamics. The control law employs an output feedback neural network adaptive element which augments an existing decoupled gain scheduled linear control law in a centralized manner to expand the class of uncertainty the system can potentially suppress. The resulting closed loop system can be shown to be uniformly ultimately bounded through a Lyapunov-like stability analysis.

A Closed-Loop Flight Control Experiment using Active Flow Control Actuators

Closed-loop pitch control on a moving 1-DOF wing model is investigated in wind tunnel experiments. The models attitude is controlled over a broad range of angles of attack when the baseline flow is fully attached using bi-directional pitching moment that is effected by flow-controlled trapped vorticity concentrations on the pressure and suction surfaces near the trailing edge. In the present work, the model is trimmed using a position feedback loop and a servomotor actuator. Once the model is trimmed, the position feedback loop is opened and the servomotor acts like an inner loop control to alter the dynamic characteristics and to introduce disturbances. Position control of the model is achieved by the flow control actuation using an arbitrary reference model based adaptive outer loop controller. The control architecture employs a neural network based adaptive element that permits adaptation to both parametric uncertainty and unmodeled dynamics.

Adaptive Control for Systems with Slow Reference Models

It is difficult to achieve good tracking performance in the presence of modeling error with the use of high adaptation gain. This leads to unnecessary high frequency control effort that can excite unmodeled dynamics. This paper introduces an adaptive control architecture that allows fast adaptation for systems with slow reference models. Fast adaptation is achieved using a high bandwidth state emulator to train a neural network. Low bandwidth control action is maintained using a filter to isolate fast emulator dynamics from the control effort. The usefulness of the architecture is illustrated on a nonlinear model for wing rock and a longitudinal model for a Boeing 747.