1. UKM Learning and Research Repository
2. PHD Thesis / Tesis PHD
3. Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina

Title:	Optimization of model reference adaptive controller for four-DoF lower limb exoskeleton joints
Authors:	Mohammad Soleimani Amiri (P90554)
Supervisor:	Rizauddin Ramli, Assoc. Prof. Ir. Dr.
Keywords:	Universiti Kebangsaan Malaysia -- Dissertations Dissertations, Academic -- Malaysia Lower Limb Exoskeleton Controller Wearable robot
Issue Date:	26-Aug-2020
Description:	Lower Limb Exoskeleton (LLE) is used as a wearable robot to enhance and amplify the strength of wearer's lower limb. In addition, it is functioned for gait training and rehabilitation. The LLE requires a controller to reduce the steady state errors and improve the stability of the robot. In this study, the control system of a 4-Degree of Freedom (DoF) LLE has been developed by Initialized Model Reference Adaptive Control (IMRAC) and Proportional-Integral-Derivative (PID) controller. The LLE consists of a waist and two legs with four joints at hip and knee. Mathematical model of LLE and its actuators are determined by Lagrangian and Kirchoff's law. Parameters of the mathematical model are estimated via Genetic Algorithm (GA) to minimize the error of the simulated and actual angular trajectories of the DC motors which are used as actuators of the LLE. The mathematical model is used as the plant in the closed-loop PID control system. A hybrid of GA and Particle Swarm Optimization (GaPSo) is applied to acquire the controller's parameter for each joint for minimizing the error which is difference between actual and desired angular trajectory. In addition, Ziegler-Nichols (Z-N) and Adaptive Particle Swarm Optimization (ZaPSo) is utilized to minimize the error of the control system by determining parameters of PID controller. PID tuned by Z-N are used as the initial value for ZaPSo. The tuned PID controller is selected as initial value for IMRAC, at which adjustment mechanism is a gradient based method for real-time adaptation of tuned PID controller. ALyapunov function has been applied to confirmthe stability of IMRAC. The mathematical model of the LLE is validated and compared with a conventional method using MATLAB/SIMULINK package. The tuned PID controller using GaPSo and PSO are verified in the closed-loop control system under step response. Furthermore, the results are compared with conventional methods such as GA, PSO and Z-N. The results illustrate that the performance of the GaPSo is 19, 36, and 14 better than GA, PSO, and Z-N. In addition, ZaPSo performs 25, 43, and 20 better than GA, PSO, and Z-N, respectively. Moreover, IMRAC is validated in LLE and its results are compared with the conventional model reference adaptive controller. Numerical analysis indicates that, the average error of IMRAC for hip and knee has lower error and better performance than the MRAC. As a conclusion, the genetically estimated mathematical model of the LLE performed similarly to the actual LLE with appropriate error. This ascertained that the optimization methods GaPSo and ZaPSo performed better than the conventional methods of PID tuning. In addition, the proposed method to initialize the IMRAC performed more accurate than the conventional model reference adaptive controller.,Ph.D.
Pages:	111
Publisher:	UKM, Bangi
Appears in Collections:	Faculty of Engineering and Built Environment / Fakulti Kejuruteraan dan Alam Bina

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.