Fuzzy-PID Position Control of a Three Fingered Robot Hand for Grasping Varying Loads / Norshariza Mohd Salleh ...[et al.]

• Main Authors: Mohd Salleh, Norshariza (Author), Ahmad Shauri, Ruhizan Liza (Author), Nasir, Khairunnisa (Author), Remeli, Nurul Hanani (Author)
• Format: Article
• Language: English
• Published: Faculty of Mechanical Engineering Universiti Teknologi MARA (UiTM), 2017.
• Subjects: Fuzzy arithmetic; Fuzzy logic; TJ Mechanical engineering and machinery; Article
• Online Access: Get fulltext; View Fulltext in UiTM IR

 In previous study, PID control was used to control the joint's motor position of a three fingered robot hand. However, it can be observed that PID control alone could not cater the nonlinearities caused by varying weight loads and friction from the gears. Both the settling time and rise time increased as the weight exceeded 20g, thus deteriorated the transient response performance. Therefore, this paper discusses the development of an intelligent system with PID control namely Fuzzy-PID control for a 6-DOF robot hand. The proposed control is capable of tuning the PID gains automatically when different loads are applied. Fuzzy Inference System was established based on the previous investigation on the effects of varying loads with the transient response parameters where the settling time and rise time have been taken as inputs and PID gains as the outputs. The proposed control specifically focused on one of the 2-DOF robotic finger. The same control inputs were applied to the motors of the other two fingers to provide synchronous motion. Finally, the proposed control was verified in real-time experiments for varying weight loads from 0g to 100g with 20g increments in terms of percent of overshoot, settling time, rise time and steady state error. From the experimental results, Fuzzy-PID control proved that it is capable to improve TS by the maximum of 17.75% and the maximum of 1.21% of TR compared to PID control only. No significant improvement was observed for percentage of overshoot and steady state error. Furthermore, the proposed control has successfully applied for random weights grasping by the 6-DOF robot hand with TS and TR improvement of 14.51% for joint 1, 9.82% for joint 2 and 3.54% of joint 1 and 2.69% of joint 2 respectively compared to PID control. Hence, the proposed Fuzzy-PID can be applied for grasping objects with different weight in future work.