Abstract
To address the complexity of traditional manipulator structures, a bionic five-finger manipulator has been designed and developed. After introducing the manipulator’s structure and functionality, the kinematic model was established using the D-H parameter method, followed by the analysis of forward and inverse kinematics. The fingertip pose was derived using the homogeneous transformation formula, and the accuracy of the theoretical analysis was validated by comparing the mathematical and MATLAB simulation results. The Monte Carlo method was employed to determine the accessible motion space of the mechanical fingertip. Subsequently, trajectory planning for the manipulator was performed based on known initial and end finger angles. A comparison with ADAMS simulation results demonstrated the rationality of the mechanical finger’s movement space, stable speed, and compliance with actual grasping requirements. This study lays a theoretical foundation for the subsequent parameter optimization of the manipulator.
Introduction
Designing robotic hands that mimic human dexterity remains a significant engineering challenge. Current designs often struggle with a trade-off between mechanical simplicity and functional precision. Bulky systems require complex control, while lightweight rope-driven models often suffer from inaccuracies under tension. This research introduces a 15-degree-of-freedom (DOF) bionic manipulator designed to bridge this gap, offering independent control over joints to facilitate complex "pinching and enveloping" maneuvers.
Design and Methodology
The manipulator features a hybrid actuation system. Electric "putting rods" drive the proximal and middle phalanges for reliable transmission, while micro-motors (steering engines) control the distal phalanges for high-precision fingertip movement. A key design feature is the thumb's 35-degree lateral tilt, which enables the transition from a flat palm to an "enveloping" grip necessary for irregular objects.
The validation process involved three main stages:
- Kinematic Modeling: The Denavit-Hartenberg (D-H) parameter method was used to map joint angles to fingertip positions in 3D space.
- Workspace Analysis: The Monte Carlo method generated a point cloud of thousands of joint configurations to define the hand's total reachable volume.
- Simulation: MATLAB's Robotics Toolbox verified the mathematical model, while ADAMS software tested the multibody dynamics to ensure smooth, vibration-free movement.
Key Results
The study confirmed high mathematical precision, with an error of less than 0.001% between theoretical models and MATLAB simulations. The hand's workspace was quantified within a range of -50 to 100mm (X), -110 to 110mm (Y), and -13 to 17mm (Z), confirming its ability to handle various object sizes. Furthermore, ADAMS simulations showed continuous and smooth velocity curves, indicating stable grasping capabilities without mechanical jerks.
Conclusion
This research provides a robust theoretical and simulation-based foundation for a 15-DOF bionic hand. While the current results are limited to virtual environments, they demonstrate that the D-H model and hybrid actuation design are viable for sophisticated robotic tasks. Future work will focus on physical prototyping to account for real-world factors like material friction and sensor noise.
Summary updated