| Kurzfassung: |
For many decades, the scientific community has devoted a tremendous amount of attention to the design of efficient controllers for nonlinear plants, which can be problematic because of challenges related to real world applications. However, there is still a need for addressing some issues that remain not clearly solved related to limitations imposed by some assumptions or some complicated approaches hard to apply for practitioners. There is also a need for approaches able to handle simultaneously several potential issues without any unnecessary additional energy usage and with improved control performances. Therefore, this book presents the design of efficient adaptive nonlinear controllers for strict-feedback nonlinear systems by taking into account multiple challenges for increased safety and reliability. The considered challenges are external disturbances, uncertain dynamics, actuation faults, unmeasured states, constrained input, unknown control direction, and singularity in the control law. Some adaptive nonlinear control schemes, which are relatively easy to apply for practitioners in order to tackle simultaneously and efficiently some and/or all aforementioned issues are presented. The book presents the design of schemes based on a reaching law-based Sliding Mode Control strategies, combined with the Input-Output Feedback Linearization technique. These schemes use Radial Basis Function Neural Networks (RBFNN) and Fuzzy Neural Networks (FNNs) to approximate the unknown system dynamics. It is illustrated how a model-free high-gain state observer is employed for estimating the unavailable system state variables. For dealing with the unknown control direction, a Nussbaum type function is presented. The schemes presented in this book have a wide scope of potential applications as they overcome important restrictions imposed by some assumptions found in many works, while ensuring very good transient and steady state performances (no peak overshoot, shorter settling time, smaller error bound and Root-Mean-Square-Error) with low leveled continuous control efforts. These canceled restrictions are the requirements about the knowledge of bounds for system dynamics uncertainties, for RBFNN or FNN approximation errors, for actuator’s faults and for external disturbances, the knowledge of control direction, the availability of full-state measurement, and the requirement about the positive-definiteness of the control gain function with known lower and upper bounds. Nonlinear controllers are designed for some particular engineering applications like - the Boeing 747-100/200 pitch angle of attack control, the tracking control of an inverted pendulum
- a 3D chaotic system synchronization
- the tracking control of a one-link manipulator with DC motor
- the control of an Unmanned Surface Vehicle Steering System
- the control of a 5 DOF upper-limb exoskeleton robot for assisted rehabilitation therapy
- the control of a disturbed Unmanned Aerial Vehicle (UAV)
- the control of a 3-axis MEMS Gyroscope
Simulation results for these application examples are provided with the corresponding MATLAB codes/SIMULINK models and compared with those reported in the literature where other adaptive schemes have been applied to the same systems so that the reader can easily see the validation of the control schemes presented in this book. |