Improved Assistive Profile Tracking of Exosuit by Considering Adaptive Stiffness Model and Body Movement

Area of Science:

• Robotics
• Human-Computer Interaction
• Biomechanics

Background:

• Exosuits offer lightweight, soft wearable robot assistance for human physical performance.
• Controlling assistive force in soft exosuits is challenging due to material properties.

Purpose of the Study:

• To enhance the performance of basic admittance-based force controllers for exosuits.
• To improve accuracy and reduce delays in assistive force generation.

Main Methods:

• Method A: Cable control synchronized with thigh motion to minimize force generation delays.
• Method B: Phase-based division of the stiffness feedforward model for accurate nonlinear stiffness compensation.
• Method C: Real-time optimization of the stiffness feedforward model using an adaptive moment estimation optimizer.

Main Results:

• Controller III, integrating all three proposed methods, demonstrated superior performance.
• Assistive force root-mean-square error reduced by 65.56% (from 39.84 N to 13.72 N).
• Time delay in force generation decreased by 65.87% within the gait cycle (from 9.99% to 3.41%).

Conclusions:

• The combined controller significantly improves assistive force accuracy and responsiveness in soft exosuits.
• The proposed methods offer a viable solution for enhancing human-robot physical interaction.
• Further development in exosuit control can lead to more effective physical assistance technologies.