| dc.description.abstract | Conventional probe endoscopy is currently used to diagnose and treat the gastrointestinal tract. However, due to its high stiffness, low flexibility, and bulkiness, it could cause pain and discomfort to the patients, when placed inside their bodies. In contrast, a wireless robotic capsule endoscope is the newly discovered diagnostic tool to comfortably investigate the entire gastrointestinal tract with no requirement for sedation, but its commercial usage is limited due to its passive motion. Hence, bioengineers and endoscopists are significantly concentrated on the controlled actuation of the wireless robotic capsule endoscope. Research in endoscopy is crucial for diagnosing gastrointestinal diseases. A robotic capsule endoscope has been evolving to examine the gastrointestinal tract, which is expected to replace conventional probe endoscopy. Literature relating to the miniature robotic capsule endoscopes is reviewed in this research, in which individual propulsion technique offers various advantages and disadvantages. However, some difficulties still need to be solved for these robotic capsule endoscopes to be used universally and commercially. Consequently, this led to the motivation to carry out this research, which could help the robotic capsule endoscope to be implemented in in-vivo medical applications. The main objectives of this research involve modelling and actively controlling the robotic capsule endoscope. This research studied the trajectory tracking control of the robotic capsule endoscope from the viewpoint of an underactuated dynamic system. A seven-stage movement plan of the robotic capsule endoscope from the literature is offered, to outline a trajectory profile for the robotic capsule endoscope to move successfully within the desired direction. Initially, the proposed dynamic model of the robotic capsule endoscope included a capsule shell, an inner mass, and a piezoelectric actuator. Next, an electromagnetic motor is used as a replacement for the piezoelectric actuator, to actively control the robotic capsule endoscope as an internal propulsion-based approach. The dynamics of the proposed control systems are then mathematically modelled and later expressed in the state-space representation, to carry out the simulations for obtaining the position and velocity trajectories of the robotic capsule endoscope. The incompleteness of the interaction model between the gastrointestinal tract and the robotic capsule endoscope is one of the governing causes of the current limitations of the robotic capsule endoscope, such as the poor understanding of involuntary examinations, and the low controllability. Subsequently, an interaction model from the literature is implemented in this research, to consider the complexity and flexibility of the gastrointestinal tract lumen. This model is applied based on the nonlinear viscoelasticity, where the interaction force involves, viscous resistance, Coulomb friction, and environmental resistance. Robust control techniques are required to obtain a high level of accuracy and precision within a more flexible and reliable system, to converge the trajectories of the states into the system within a finite time. This research designs various control strategies, including Open-Loop Control, Closed-Loop Control, Linear Quadratic Regulator, and Sliding Mode Control and its variations involving first-order Sliding Mode Control, second-order Sliding Mode Control, Integral Sliding Mode Control and Terminal Sliding Mode Control. MATrix LABoratory (MATLAB) software is then used to simulate the controllers and optimise their performance. Proportional-Integral-Derivative is a conventional control method, which is modified and used as a Proportional-Derivative format in this research. Linear Quadratic Regulator is an exclusively linear control method applied for linear systems, while Sliding Mode Control is a robust control method used for complex systems where disturbances and uncertainties are present. Accordingly, linearisation is performed in this research to linearise the model, due to the nonlinearity of the robotic capsule endoscope system. Additionally, the Lyapunov function is used to confirm the stability of the designed control techniques. In conclusion, the theoretical analysis and simulation results obtained from the SIMULINK and SIMSCAPE Multibody models successfully verified the proposed position and velocity profiles and validated the trajectory control algorithms of LQR and SMC. SIMSCAPE Multibody is then further implemented to visualise the robotic capsule endoscope trajectory and consider the impact of strain conditioning on the designed RCE models. The simulation results from Sliding Mode Control are discovered to be more accurate, than the Linear Quadratic Regulator and Proportional-Derivative control methods. Nevertheless, there is a delay in following the desired trajectory, which could be revised in future work. Moreover, as the future plan of this research is to construct and experiment the RCE, its fabrication and experimentation processes are also discussed in this study. | en_US |
