Autonomous slip control inspired by human physiology for improved shared control strategy

Area of Science:

• Robotics
• Biomechanics
• Neuroprosthetics

Background:

• Replicating human hand functionality in upper-limb prostheses is challenging due to limited user control inputs and sensory feedback.
• Existing semi-autonomous control strategies aim to improve prosthesis dexterity by integrating sensing technologies.
• Myoelectric control limitations necessitate advanced solutions for effective prosthetic hand operation.

Purpose of the Study:

• To develop an improved low-level controller for prosthetic hands focused on slip detection and autonomous grasping.
• To enhance object holding stability in artificial hands by refining grasp control strategies.
• To reduce the reliance on user input for maintaining a stable grasp during object manipulation.

Main Methods:

• A shared control strategy dividing grasp control into user-initiated high-level and stability-focused low-level controllers.
• Implementation of a novel slip module utilizing distributed 3D force sensors and a friction cone strategy.
• Bandpass filtering for establishing an initial stable grasp model without prior knowledge.

Main Results:

• The proposed controller demonstrated efficacy in preventing object slippage during grasping tasks.
• The system required less grasping force compared to conventional state-of-the-art approaches.
• Qualitative validation showed positive responsiveness to unexpected weight changes, comparable to human grasping.

Conclusions:

• The novel slip detection module enhances autonomous grasping capabilities in prosthetic hands.
• This approach offers a more stable and efficient grasp, reducing the risk of object slippage.
• The findings contribute to the development of more functional and intuitive upper-limb prostheses.