Multi-phase optimisation model predicts manual lifting motions with less reliance on experiment-based posture data.

Area of Science:

• Biomechanics
• Ergonomics
• Human Motion Analysis

Background:

• Optimisation-based predictive models are crucial for understanding lifting strategies.
• Current models often depend on empirical, subject-specific posture constraints, limiting their generalizability.
• Over-reliance on these constraints restricts the wider application of predictive biomechanical models.

Purpose of the Study:

• To introduce a novel multi-phase optimisation method (MPOM) for predicting two-dimensional, sagittally symmetric semi-squat lifting.
• To enhance prediction accuracy while reducing the dependency on experimental data and subject-specific constraints.
• To decompose the lifting task into distinct phases for more refined prediction.

Main Methods:

• The proposed multi-phase optimisation method (MPOM) divides the lifting task into three phases: initial posture, final posture, and dynamic lifting.
• Force and stability strategies are employed for predicting the initial and final postures.
• A smoothing-related objective is utilized for the dynamic lifting phase.
• Validation involved collecting box-lifting motions with varying initial box heights.

Main Results:

• MPOM demonstrated comparable or superior accuracy to traditional single-phase optimisation (SPOM) methods, such as minimum muscular utilisation ratio.
• The MPOM approach significantly reduces the need for extensive experimental data.
• The method offers potential for improved accuracy, contingent on careful weighting in posture prediction phases.

Conclusions:

• The multi-phase optimisation method (MPOM) provides an effective alternative for predicting human lifting motions.
• MPOM enhances prediction accuracy and reduces reliance on restrictive, subject-specific constraints.
• This approach broadens the applicability of optimisation-based models in biomechanics and ergonomics.