Online Estimation of Elbow Joint Angle Using Upper Arm Acceleration: A Movement Partitioning Approach

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Neuroscience

Background:

• Functional Electrical Stimulation (FES) systems require accurate joint angle estimation for effective control.
• Elbow-shoulder synergies are crucial for understanding upper limb movement in individuals with spinal cord injury (SCI).
• Assisting C5/C6 SCI patients necessitates reliable methods for estimating elbow joint movement.

Purpose of the Study:

• To develop an online method for estimating elbow flexion/extension angles using upper arm acceleration signals.
• To implement a hierarchical approach incorporating 'movement phases' for structured and efficient angle estimation.
• To enhance FES control systems through accurate and real-time joint angle prediction.

Main Methods:

• A three-level hierarchical structure involving clustering, Hidden Markov Models (HMMs) for recognition, and neural networks for angle estimation.
• Partitioning of Activities of Daily Living (ADLs) into distinct movement phases.
• Utilizing upper arm acceleration data and incorporating varied object locations to improve neural network generalization.

Main Results:

• Achieved a cross-correlation coefficient (K) of 90.25% and Normalized Root Mean Squared Error (NRMSE) of 13.64% for estimated elbow angles.
• A post-processing technique improved results to K = 91.19% and NRMSE = 12.83%, addressing discontinuity intervals.
• Demonstrated the effectiveness of the 'movement phases' approach for online estimation.

Conclusions:

• The proposed online estimation method is valuable for FES control systems, particularly for C5/C6 SCI patients.
• The hierarchical structure and 'movement phases' approach provide an efficient and structured method for joint angle estimation.
• The study highlights the potential of using upper arm acceleration for real-time elbow angle tracking in rehabilitation applications.