A Model of Multi-Finger Coordination in Keystroke Movement

Area of Science:

• Biomechanics
• Neuroscience
• Robotics

Background:

• Professional pianists exhibit precise, coordinated multi-finger keystroke movements regulated by motor neural centers.
• Current research on piano keystroke coordination lacks accurate modeling, focusing on theoretical descriptions.
• The ring finger's limited flexibility due to tendon connections necessitates specific study in multi-finger coordination.

Purpose of the Study:

• To develop a predictive model for coordinated keystroke actions, focusing on the middle and ring fingers.
• To establish a practical modeling method for multi-finger coordination in piano playing.
• To provide scientific guidance for piano learners and inform motion planning for exoskeleton robots.

Main Methods:

• A motion measurement platform utilizing Leap Motion collected data from 12 professional pianists.
• A backpropagation (BP) neural network model was developed for multi-finger coordination prediction.
• The BP model was optimized using a genetic algorithm (GA) and a sparrow search algorithm (SSA).

Main Results:

• The sparrow search algorithm-optimized backpropagation (SSA-BP) neural network model achieved high predictive accuracy.
• The model's root mean square error was 4.8328° for predicting the ring finger's MCP joint angular rotation.
• Keystroke motion of the ring finger's MCP joint was accurately predicted using middle finger motion data and individual differences.

Conclusions:

• The SSA-BP model offers an accurate method for predicting ring finger keystroke motion based on middle finger data.
• This predictive capability provides a valuable evaluative tool for training piano learners in multi-finger coordination.
• The findings contribute to understanding and modeling complex human motor skills for applications in music and robotics.