Cardiff University | Prifysgol Caerdydd ORCA
Online Research @ Cardiff 
WelshClear Cookie - decide language by browser settings

Optimization of non-linear robust controller for complex multi-link robotic system

Abdul Samad, Bdereddin ORCID: 2023. Optimization of non-linear robust controller for complex multi-link robotic system. PhD Thesis, Cardiff University.
Item availability restricted.

[thumbnail of Optimization of Non-Linear Robust Controller for Complex Multi-link Robotic System.pdf] PDF - Accepted Post-Print Version
Restricted to Repository staff only until 19 July 2024 due to copyright restrictions.

Download (5MB)
[thumbnail of Cardiff University Electronic Publication Form] PDF (Cardiff University Electronic Publication Form) - Supplemental Material
Restricted to Repository staff only

Download (234kB)


This research focuses on integrating artificial intelligence and knowledge-based systems to improve the control of a complex multi-link mechanism. The Robogymnast is developed and analysed as a platform to study the intricacies and challenges of a three-link robot system. Through modelling, simulation, and advanced control techniques, the study aims to enhance the overall performance and manoeuvrability of underactuated mechanisms, contributing to advancements in robotics. The linearized mathematical model is employed to explore state space determination in the system. The motion of the robot is represented mathematically using Lagrange equations. However, controlling the movements of the robot gymnast poses challenges due to its nonlinear and multivariate characteristics. The proposed approach for controlling the three-link Robogymnast robotic gymnast and evaluating its stability is examined and compared with existing methods. It compares the effectiveness of a conventionally configured linear quadratic regulator (LQR) with a hybrid approach that combines fuzzy logic and LQR (FLQR) for stabilizing the Robogymnast. The study investigatesthe application of LQR and FLQR controllers to the Robogymnast, analysing the system's behaviour in five scenarios, including the original value and distributions of ±25% and ±50%. It also explores factors affecting swing-up control in the underactuated three-link Robogymnast. Additionally, a system simulation using MATLAB Simulink is conducted to demonstrate the impact of factors such as under/overshoot, rise time, and settling time. A linear quadratic regulator/fuzzy logic controller is employed to stabilize a three-link robotic mechanism. The controller system is optimized using two algorithms: Teaching Learning-Based Optimization (TLBO) and Particle Swarm Optimization (PSO). The results demonstrate that the TLBO algorithm significantly enhances system stability compared to the conventional PSO algorithm. Specifically, the TLBO algorithm achieves a reduction in the overshoot metric to zero for the first link, 39% for the second link, and 23% for the third link. Moreover, the TLBO algorithm exhibits shorter rising and settling times. Notably, the Integral of Time multiplied by Absolute Error (ITAE) for the first joint is 1.688 with the TLBO algorithm, while it is 2.68 with the PSO algorithm. The ITAE values for the second and third links are approximately 0.3117 and 0.02145, respectively, for both algorithms. Lastly, a new approach is developed to control the movement of the pendulum system through synchronization, and the performance of the system is investigated using the Robogymnast at Cardiff University. A simulation is created using MATLAB/Simulink to study the system's motion and swinging-up behaviour. The simulation of the Robogymnast and the implementation of the controllers are carried out using MATLAB® and STM32 microcontroller in the C++ program environment, respectively. The similarity of joints' motion in the real system and simulation exhibits error percentages of 30% or less, indicating reliable and accurate results for these joints. The research provides valuable insights into the optimization and design of robotic systems using advanced control techniques and optimization algorithms.

Item Type: Thesis (PhD)
Date Type: Completion
Status: Unpublished
Schools: Engineering
Uncontrolled Keywords: 1) Robogymnast 2) Triple-link Robotic system 3) FLQR Control 4) LQR Control 5) TLBO Algorithm 6) PSO Algorithm
Date of First Compliant Deposit: 19 July 2023
Last Modified: 16 Aug 2023 09:37

Actions (repository staff only)

Edit Item Edit Item


Downloads per month over past year

View more statistics