A New Framework of Model Reference Adaptive Control–Partial-State Feedback Designs and Applications

Model reference adaptive control (MRAC) is an important methodology to accommodate various system uncertainties. Traditionally, for output tracking MRAC, either a state feedback controller or an output feedback controller is used. As an effort to provide additional feedback capacity and design flexibility to the existing MRAC family, this dissertation focuses on the development of partial-state feedback MRAC framework and the application of such novel MRAC designs. For partial-state feedback MRAC, plant-model matching is achievable as with full-state feedback control, while the controller structure enjoys less complexity as compared with an output feedback MRAC design. In this study, adaptive partial-state feedback control designs are developed for single-input-single-output systems and multi-input-multi-output systems, respectively. Both adaptive control schemes ensure closed-loop system stability and asymptotic output tracking. Related issues such as plant-model matching, error model, adaptive law, and stability analysis are investigated in this dissertation.

Applications of the new partial-state feedback MRAC designs are also explored. Based on the enhanced robustness brought by the partial-state feedback MRAC designs, new sensor failure compensation control schemes for single-input-single-output systems and multi-input-multi-output systems are developed and investigated in this dissertation. The new adaptive sensor failure compensation schemes have the capability of ensuring asymptotic output tracking while compensating all possible uncertain sensor failures in the presence of the system parametric uncertainties. The new sensor-redundancy-free compensation schemes relax some requirements of traditional fault-tolerant control techniques. This dissertation also extends the partial-state feedback MRAC application to the multi-agent system control area. A new adaptive output consensus control scheme via partial-state feedback MRAC is developed which can increase control design flexibility and make full use of all possible system measurements for multi-agent consensus control. The new consensus control scheme is able to achieve closed-loop signal boundedness and asymptotic output consensus in the presence of system parameter uncertainties.

The effectiveness of the developed adaptive partial-state feedback control designs and sensor fault compensation designs have been assessed on some high-fidelity aircraft systems or quadrotor systems by MATLAB. The simulation results have demonstrated the desired performance of our designs.