PID Control of Proximal and Distal Interphalangeal Jointed Robotic Hand

Abstract

This study presents the design, modeling, and control of a biomimetic robotic hand with independently actuated proximal (PIP) and distal (DIP) interphalangeal joints. A detailed anatomical reconstruction of the human hand was created in a CAD environment, where all finger joints were modeled as single-degree-of-freedom rotary joints. A total of 14 MG90S micro servo motors—three for each of the four fingers and two for the thumb—were integrated into the design to achieve a fully actuated, multi-input–multi-output (MIMO) structure capable of independent joint control.

Following the mechanical design, the model was transferred to a Python-based simulation environment, and PID controllers were implemented for MCP, PIP, and DIP joints. PID parameters were tuned and compared using Ziegler–Nichols and Cohen–Coon methods. Closed-loop angular responses of each joint were analyzed with respect to rise time, overshoot, damping ratio, and settling performance. The results show that while Ziegler–Nichols tuning provides rapid response, it introduces significant overshoot and oscillatory behavior, particularly in low-inertia joints such as the DIP. Conversely, the Cohen–Coon method yields more balanced, stable, and well-damped responses across all joints, making it a more suitable choice for robotic finger control where precise manipulation and stability are required.

This study demonstrates that a fully actuated robotic hand that adheres to anthropometric joint structures can successfully achieve natural finger motion using PID-based independent joint control. The findings provide an important foundation for future applications in prosthetics, rehabilitation robotics, and dexterous robotic manipulation.