Adaptive control of mechanical systems
Date of Issue2014
School of Electrical and Electronic Engineering
Robotics Research Centre
Nonholonomic underactuated mechanical systems are special systems full of research interest and practical sense. Referring to those mechanical systems whose number of control variables is less then the degrees of freedom, the underactuated systems are abundant in real life, ranging from landing vehicles, surface ships, underwater vehicles to spacecrafts. Mobile robots and surface vessels are two typical nonholonomic underactuated systems that receive tremendous consideration in literature. For the tracking and stabilization control of underactuated mechanical systems, many methodologies have been proposed for controlling these systems by researchers. However, there are still a lot of open problems in this area. Among them the tracking control with input saturation and output-feedback control in the presence of parametric uncertainties are two important problems to be solved in this thesis. For tracking and stabilization control of nonholonomic mobile robots with input saturation, a new adaptive scheme is proposed to ensure that the bounds of the control torques are functions of only design parameters and reference trajectory. Thus suitable design parameters can be chosen so that such bounds are within the given saturation limits. For adaptive output-feedback tracking control of nonholonomic mobile robots, a new adaptive control scheme is proposed including designing a new adaptive state feedback controller and two high-gain observers to estimate the unknown linear and angular velocities respectively. For tracking control of underactuated ships we design the yaw axis torque in such a way that its corresponding subsystem is finite time stable, which makes it be decoupled from the second subsystem after a finite time. This enables us to design the torque in the surge axis independently. For adaptive output-feedback tracking control of underactuated ships, by using the prescribed performance bound technique, the position error and orientation error can be guaranteed converging to arbitrarily small residual sets at a pre-specified exponential rate. Distributed coordination for a group of dynamic agents has attracted many researchers in recent years, such as cooperative control of unmanned air vehicles (UAVs), formation control, ocking control, distributed sensor networks, attitude alignment for a clusters of satellites, congestion control in communication networks etc. An important performance indicator for the consensus problem is the convergence rate. Most of the existing consensus control schemes for multi-agent systems so far achieve asymptotical convergence. This implies that the convergence rate at best is exponential, and it needs infinite time for the tracking errors to converge to the origin. Thus finite-time consensus control of a group of mechanical systems with parametric uncertainties is an important problem and will be considered in this thesis. New adaptive finite time continuous distributed control algorithms are proposed for multi-agent systems under two different scenarios. For the leaderless multi-agent systems, it is shown that the states of the mechanical systems can reach a consensus within finite time. For the leader-follower control of a group of mechanical systems, two control schemes are proposed. The first one is based on the hierarchical decomposition of the graph. It is demonstrated that perfect consensus can be reached or the leader-following errors converge to a region with arbitrarily small radius in finite time. The second one is based on distributed estimators, which also achieves finite-time consensus. Transient performances in terms of convergence rates and time are also analyzed. Finally simulation results illustrate and verify the effectiveness of the proposed schemes.
DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation