On Network-based Control and Sensitivity Characterization of Mobile Robot in Intelligent Space

Abstract

This dissertation addresses the problem of path-tracking control of a mobile robot, also called an Unmanned Ground Vehicle (UGV), in Intelligent Space, where the controller is located on an entity different from the robot and communicates with the robot over a communication network. The involvement of a communication network leads us to the core of this research, the network time-delay factor. The existence of a network delay presents a challenging problem that might degrade the overall system performance and even destabilize the closed-loop control system.

The existing research area for the aforementioned scenario is called Network-based control system (NBC) mostly focused on a general linear system for which the controller must be redesigned so that the overall NBC system can work properly. Distinct from the existing research, and innovative in its own right, the research presented in this dissertation focuses on a specific nonlinear system, the remote UGV path-tracking. More specifically, we focus on the methods that allow the existing workable path-tracking controller to be reused in the NBC environment.

In this work, Accumulated effect parameter tuning method is firstly proposed to tune the geometrical path-tracking controller used in UGV before operating over communication network; then sensitivity analysis is introduced to consider how the system is sensitive to noise or perturbation so that the operating condition, such as UGV speed and path curvature, may be changed to limit the effect from noise or perturbation; afterwards, Feedback preprocessor (FP) is proposed to alleviate the effect of network delay by using UGV position estimation through UGV kinematics model; along with FP, UGV response time is proposed to demonstrate the effect of different UGV characteristics on path-tracking performance; finally, the effect of using Gain scheduler (GS) with two-dimensional and one-dimensional gain table is investigated for the capability to alleviate the network delay on remote UGV path-tracking.