Harvard SEAS Pioneers New Method for Designing Advanced Robotic Joints

Researchers at the Harvard John A. Paulson School of Engineering and Applied Sciences (SEAS) have developed a new method for designing knee-like joints in robots. These joints, termed rolling contact joints, aim to enhance robotic grippers, improve assistive devices for humans, and enable more fluid movement in robots.

Design Approach

The new method optimizes rolling joints using computer-based adjustments to the shape of each joint component. This process allows the design to be tailored to a specific force requirement or application, such as a robotic gripper or a human-like robot's appendage.

This approach is intended to result in more efficient robots, potentially allowing for smaller actuators by directing energy to specific operational needs. Robert J. Wood, a professor at SEAS, noted:

"The goal is to integrate motion control into the robot's mechanics and materials, reducing the control system's workload to focus on task-level objectives."

Development Inspiration

The concept for this joint design method originated from a project to create a soft robotic gripper capable of both gentle wrapping and strong force application. This involved examining how to combine rigid links with flexible joints, similar to the bone and cartilage structure of a human hand, leading to a focus on rolling contact joints.

These joints consist of curved surfaces that roll against each other, connected by flexible elements. While traditional robotics often use software and control algorithms for joint movement, this new method integrates these choices into the joint's geometric design.

Advantages and Prototypes

Rolling contact joints offer advantages such as flexibility, low friction, and high wear resistance, differentiating them from more common bearings and four-bar linkages. The Harvard team's mathematical method allows for the creation of noncircular and irregular joint shapes, contrasting with traditional rolling contact joints that use circular surfaces.

To validate their method, the team constructed two prototypes:

• Knee-like Joint: This prototype, designed to mimic the average path of a human knee, demonstrated a 99% reduction in misalignment compared to standard devices. This outcome suggests potential applications in tailored knee braces, exoskeletons, and joint replacements.
• Robotic Gripper: A two-finger gripper prototype was optimized to deliver maximum force based on object size. This gripper was capable of holding over three times the weight of a standard version using the same actuator input.

Future Implications

This capability to optimize human-like joints for various applications is expected to open new areas of research, including the development of task-specific robots, advanced assistive robotics, and studies in animal biomechanics.