Dynamic Balance Gait for Walking Assistance Exoskeleton.

Area of Science:

• Robotics
• Biomechanics
• Rehabilitation Engineering

Background:

• Powered lower-limb exoskeletons offer mobility for spinal cord injury (SCI) patients.
• Current exoskeletons often use predefined gaits, leading to unnatural movement and instability.
• Existing systems struggle to adapt to disturbances during walking.

Purpose of the Study:

• To develop a novel gait planning approach for powered lower-limb exoskeletons.
• To provide reliable and balanced walking assistance for SCI patients.
• To enhance exoskeleton adaptability to external disturbances.

Main Methods:

• Modeled the patient-exoskeleton system as a linear inverted pendulum (LIP).
• Utilized an orbital energy diagram to estimate patient intention.
• Employed dynamic movement primitives (DMP) for dynamic gait trajectory planning.
• Dynamically updated DMP parameters to improve disturbance rejection.

Main Results:

• Validated the proposed approach in a human-exoskeleton simulation platform.
• Demonstrated the effectiveness and advantages of the novel gait planning method.
• The approach successfully generated reliable and balanced gaits.

Conclusions:

• Decomposed dynamic balance gait into intention estimation, discrete planning, and continuous regulation.
• An intention estimator uses sensory information to predict stepping.
• A discrete gait planner generates trajectories based on step parameters (S, T).
• A continuous gait regulator adjusts gait to counteract disturbances during walking.