Ahmad Ansari

Adaptive control of an aircraft with uncertain nonminimum-phase dynamics

This paper investigates four control architectures that use adaptive control to follow step, ramp, and harmonic roll-angle commands for a linearized aircraft model with an unknown transition from minimum-phase to nonminimum-phase (NMP) dynamics. In particular, we consider retrospective cost adaptive control (RCAC) with 1) a command-feedforward control architecture; 2) an output-feedback control architecture; 3) a centralized control architecture that uses both command feedforward and output feedback; and 4) a decentralized control architecture that uses both command feedforward and output feedback. For baseline tests, we assume that the location of the NMP zero is known. The goal of this work is to improve the transient response and rate of convergence. Numerical testing shows that RCAC with the decentralized control architecture using command feedforward and output feedback gives the fastest convergence. Furthermore, resetting controller coefficients at the start of the transition improves the transient response.

Retrospective Cost Adaptive Control of Generic Transport Model Under Uncertainty and Failure

This paper presents a case study on the application of retrospective cost adaptive control to the NASA generic transport model under conditions of uncertainty and failure. To apply retrospective cost adaptive control to the generic transport model, a collection of control architectures is defined, where each architecture has a decentralized adaptation structure in the sense that a performance (error) variable is assigned to each control channel for online decentralized adaptation. Sensor signals may be shared among the channels to account for channel coupling. Then, a series of cases that examine the performance of retrospective cost adaptive control under various conditions is considered, including trim finding, startup, unknown time-varying aircraft dynamics (such as icing and mass variation), unknown flight envelope (including the possibility of conflicting commands), known sensing constraints (a failed air data system), known actuation constraints (requiring differential thrust), unknown sensor failur...

Exploration and mapping of an unknown flight envelope

The flight envelope of an aircraft consists of the constant trim states that the aircraft can attain, given in terms of airspeed, turn rate, and flight path angle. Flight trajectories typically consist of a sequence of trim commands with intermediate transitions. While the flight envelope of an aircraft is determined beforehand, it may change under off-nominal conditions due to damage or actuator failure. The goal of this paper is to investigate the ability of an adaptive control law to reach new trim states in the case where the flight envelope is totally unknown. Within simulation, this approach provides an alternative technique for mapping the flight envelope. For an aircraft in flight, this approach can be used to reach new trim states under envelope uncertainty, as may occur during off-nominal flight conditions.

Aircraft sensor fault detection using state and input estimation

This paper presents a method for detecting aircraft sensor faults using state and input estimation. We formulate the kinematics as a nonlinear state space system, which requires no modeling information, and thus is applicable to all aircraft. To illustrate the method, we investigate three fault-detection scenarios, namely, faulty pitot tube, angle-of-attack sensor, and accelerometers. We use the extended Kalman filter for pitot-tube and angle-of-attack sensor fault detection, and retrospective cost input estimation for accelerometer fault detection. For numerical illustration, we use the NASA Generic Transport Model to detect stuck, bias, drift, and deadzone sensor faults.

Adaptive trim and trajectory following for a tilt-rotor tricopter

We apply adaptive control to an unconventional aircraft, namely, a three-rotor flight vehicle, one of whose rotors can tilt about the longitudinal axis of the fuselage. This combination of actuators has aerodynamic advantages but also poses challenges in terms of trimming the aircraft in order to balance the torque about the roll, pitch, and yaw axes. The paper uses retrospective cost adaptive control (RCAC) to trim the aircraft in hover as well as to follow straight-line and circular flight trajectories.

Adaptive input estimation for nonminimum-phase discrete-time systems

The accuracy of state estimation can be enhanced by simultaneously estimating unknown inputs. This paper presents an extension of retrospective cost input estimation (RCIE) that directly updates the estimates of all states. We show that RCIE can be used for systems in which the transmission zeros from the estimated input to the measurement are nonminimum phase. We demonstrate this ability on numerical examples, and we compare the estimates from RCIE to estimates from prior methods for input estimation. Finally, we use this technique to estimate the acceleration of a flight vehicle using camera data, and we assess the accuracy of the acceleration estimates by transforming the onboard body-frame acceleration measurements to the camera frame.

Retrospective Cost Adaptive Control of the Generic Transport Model under Abrupt Faults

Retrospective cost adaptive control (RCAC) was demonstrated in using the NASA GTM under conditions of uncertainty and failure. In the present paper the goal is to stress RCAC by considering faults that are abrupt, severe, unknown, and unforeseen. In particular, we consider 1) abrupt changes in the aircraft mass properties; 2) abrupt changes in the aerodynamic coefficients; 3) abrupt control-surface deflection and jam; and 4) abrupt thrust failure. Particular attention is paid to the severity of the transient response between the abrupt failure and recovery of the aircraft.

Retrospective Cost Adaptive PID Control of Quadcopter/Fixed-Wing Mode Transition in a VTOL Aircraft

Although novel aircraft designs can be quickly conceived and built, the ability to develop and test these concepts is impeded by the inability to rapidly prototype autopilots that can accommodate unconventional airframes with nonstandard sensor/actuator configurations. For vehicles that are complex or expensive and that are intended for military or civil certification and production, autopilots are traditionally designed based on a combination of multi-body modeling, computational fluid dynamics, and wind tunnel testing. However, detailed models are not available for experimental flight vehicles, and thus manual autopilot tuning is needed without guarantees of success. To overcome this problem, the present paper focuses on an adaptive control technique that requires minimal modeling and limited tuning. In particular, we apply retrospective cost adaptive (RCAC) control to three aircraft simulation models. The first aircraft is a quadcopter, and the second is a fixed-wing vehicle. Next, we consider a VTOL vehicle that can fly in both quadcopter and fixed-wing modes. The RCAC tunings for the VTOL aircraft in quadcopter and fixed-wing mode are chosen to be identical to the tunings of the quadcopter and fixed-wing aircraft despite the differences in vehicle geometry. The goal is to assess the ability of RCAC to control different airframes with the same tunings, while also assessing the ability to control the VTOL aircraft during mode transition.

Estimation of angular velocity and rate-gyro noise for sensor health monitoring

Based on an exact kinematics model, this paper considers two strategies aimed at diagnosing the health of a 3-axis rate gyro. In the first strategy, noisy attitude measurements are used to estimate angular velocity; comparing these estimates to the actual rate-gyro measurements provides the means for assessing the health of the rate gyro. In the second strategy, noisy attitude and angular velocity measurements are used to estimate the noise corrupting the rate-gyro measurements; analysis of the noise estimate provides an alternative means for assessing the health of the rate gyro. Both strategies are formulated as input and state estimation problems, where extended retrospective cost input estimation provides an estimate of the unknown input and the unscented Kalman filter provides the state estimate.

Adaptive non-Bayesian state estimation

This paper presents an adaptive retrospective cost state estimation (RCSE) algorithm that uses no knowledge of the statistics of the process and measurement noise. To demonstrate the method, we investigate three cases for comparison with the Kalman filter, namely, known dynamics with known noise statistics, uncertain dynamics with known noise statistics, and uncertain dynamics with uncertain noise statistics. For numerical illustration, we apply RCSE to a damped oscillator, a damped rigid body, and a lateral aircraft model.