Self-Evolving Fuzzy-Based Control of Soft Pneumatic Rotary Actuators

Physical environments in the real world are complex, unstructured, and often delicate. Pneumatic soft actuators draw inspiration from the adaptability of living organisms, enabling them to operate effectively in collaborative environments. Due to their compliant nature, the soft actuators present a solution for robots to interact safely with human co-workers and their surroundings. Multiple challenges arise while designing these actuators such as modelling, sensing, and controlling their continuously deformable structures. To address these issues, this research presents the dynamic modelling and self-evolving fuzzy-based closed-loop control for Soft Bidirectional Bending Actuator (SBBA) using visual feedback. This study discusses the hyper-elastic modelling of SBBA for bidirectional actuation, establishing a non-linear relationship between applied pressure and bending angle. Exteroceptive feedback is obtained from colour markers, which is integrated with the proposed control method to regulate the input pressure to the actuator. Experimental results demonstrate that the proposed control methodology can render the desired angle with an improvement of 22.97% in rise time, 33.61% in peak overshoot, and 30.35% in peak time, resulting in faster response and lower energy consumption.