Predicting foot placement for balance through a simple model with swing leg dynamics.

Area of Science:

• Robotics
• Biomechanics
• Control Theory

Background:

• Stepping is crucial for balance recovery against external disturbances.
• Existing models often neglect swing leg dynamics, limiting prediction accuracy.
• Minimizing actuation cost during balance recovery is a key challenge.

Purpose of the Study:

• To develop a model predicting optimal step placement and duration for bipedal balance recovery.
• To investigate the role of swing leg dynamics in minimizing energy expenditure.
• To minimize swing actuation cost during recovery from external perturbations.

Main Methods:

• Expanded the linear inverted pendulum model with an actuated linear pendulum for swing leg dynamics.
• Derived closed-form expressions for swing actuation considering pelvis and foot motion.
• Calculated optimal step position and duration to minimize swing cost.

Main Results:

• Optimal step duration was found to be constant across perturbations.
• Optimal step position was linearly proportional to the magnitude of external perturbations.
• Model predictions aligned with experimental data from human subjects.

Conclusions:

• The proposed model effectively predicts recovery foot placement in bipedal systems.
• Incorporating swing leg dynamics provides a more accurate approach to balance recovery.
• Findings offer insights into human-like balance recovery strategies.