Control methods for liquid-gas phase transition soft actuators: from simulation to implementation.

This thesis presents the study, the modeling and the control of an electrothermal soft actuator. The work begins with a review of the state of the art of different kind of soft actuators and soft robotics, highlighting existing technologies and their applications. The focus then shifts to the modeling and the control of a specific electrothermal soft actuator created by PhD Diogo Fonseca of the University of Coimbra. Then, a mathematical model of the actuator was developed using formulas derived from existing literature and experimental data. This model served as basis for designing and testing different control strategies to optimize actuator performance in a simulation environment. Four control strategies were implemented and evaluated: an On-Off controller, a Proportional-Integral (PI) controller tuned with experimental data, a PI controller tuned with simulation data derived from the developed model, an adaptive Proportional-Integral-Derivative (PID) controller with variable Kp, Ki and Kd values adjusted based on equations gathered from simulation feedback. These controllers were tested extensively in both simulated environments using MATLAB and in real-world scenarios using the physical actuator setup. Different experiments were conducted to assess the actuator's performance in terms of force generation, precision, stability, and robustness as well as to determine the most effective test for evaluating and comparing the obtained control strategies. The system was integrated with an Arduino Uno connected to a power supply in order to power the actuator and the feedback acquisition was done using pressure sensors. The control algorithms were implemented in MATLAB and Python coding was created to facilitate communication between the computer and the microcontroller. The results indicate that, while all controllers perform adequately, the adaptive PID controller demonstrates superior performance in terms of stability, precision, speed and robustness; the choice of the controller, however, can be made based on the specific application requirements. It was also found that while the speed of the actuator can be slightly controlled, it is inherently limited by the physical characteristics of the actuator. Overall, this research provides insights into the optimal control of a specific electrothermal soft actuators and highlights the potential of adaptable control strategies in enhancing actuator performance.