Spatiotemporal representation of 3D hand trajectory based on beta-elliptic models

Area of Science:

• Biomechanics
• Robotics
• Human-Computer Interaction

Background:

• Modeling hand trajectory is crucial for understanding grasping.
• Existing 2D models lack 3D applicability and detailed kinematic analysis.
• Integrating geometric and velocity information offers a more comprehensive approach.

Purpose of the Study:

• To extend the 2D beta-elliptic model to 3D for hand grasping trajectory analysis.
• To incorporate both geometric and velocity parameters in the hand movement model.
• To evaluate the performance of different beta-elliptic model variations.

Main Methods:

• Decomposition of 3D task space trajectories into elementary components.
• Modeling velocity profiles using beta functions and geometry using elliptic shapes.
• Construction of a grasping movement database and assessment of reconstruction errors (distance, curvature).

Main Results:

• The 'two tangent points' beta-elliptic model variation demonstrated superior performance.
• Average and maximum relative errors were 2.73% and 8.62% respectively.
• Maximum curvature error was 8.62% for the 'two tangent points' method.

Conclusions:

• The 3D beta-elliptic model effectively captures hand grasping trajectories.
• The 'two tangent points' method offers higher accuracy in kinematic analysis.
• This model can quantify rehabilitation or teaching process improvements in hand movement.