Adaptive Control of Linear Systems with Actuator Failures
Actuator failure compensation is an important topic of both theoretical and practical significance. Actuator failures in control systems may cause severs system performance deterioration and even catastrophic closed-loop system instability. For system safety and reliability, such actuator failures must be properly accommodated.
Actuator failures are uncertain in nature. Very often it is impossible to predict in advance which actuators many fail during system operation, when the failures occur, what types and what values of the failures are. It may also be impractical to determine the failure parameters after a failure occurs. Adaptive control, which is capable of accommodating system parametric, structural, and environmental uncertainties, is a suitable for designing actuator failure compensation schemes without explicit knowledge of the occurrences of actuator failures and the failure values.
This dissertation research is focused on adaptive control of linear time-invariant (LTI) systems with unknown actuator failures characterized by some unknown inputs being stuck at some unknown fixed values. The objective is to develop a theoretical framework for adaptive control of linear systems with plant parameter and actuator failure uncertainties, guaranteeing system stability and tracking performance.
In this work, we address several adaptive actuator failure compensation problems. We first investigate the state tracking problem for LTI systems with unknown plant dynamics and actuator failures and a state feedback design is developed. We then study the output tracking problem for minimum phase LTI systems with both un-known plant parameters and actuator failures. Several designs are presented: a state feedback design and an output feedback design for single-output systems and two output feedback designs for multi-output systems. All the four designs are based on the model reference adaptive control (MR.AC) approach. Finally, we develop an out-put feedback output tracking scheme for single-output systems based on the adaptive pole placement control (APPC) approach, which is applicable to both non minimum phase and minimum phase systems.
For each design, a parameterized controller structure is first proposed, and the design conditions for plant-model matching are studied. Under the design conditions, the proposed controller structure is effective in achieving the desired plant-model matching when implemented with matching parameters. For adaptive design with plant parameter and actuator failure uncertainties, we first derive the error equations. Based on the error equations, adaptive laws updating the unknown controller parameters are derived. It is shown in detailed analysis that, for all the developed adaptive actuator failure compensation schemes, all closed-loop signals are bounded and asymptotic tracking is guaranteed, despite the uncertainties in actuator failures and plant parameters. In all the developed adaptive schemes, a fault detection and diagnosis procedure is not needed. Each design is followed by a simulation study to illustrate the performance of the developed adaptive control scheme.
As a case study, a linearized longitudinal dynamics model of a transport aircraft is also used as the controlled plant in simulations to examine the effectiveness of several designs. The obtained simulation results verify the stability and asymptotic tracking performance of the developed adaptive actuator failure compensation schemes.