High-Fidelity, Customizable Force Sensing for the Wearable Human-Robot Interface.

Area of Science:

• Robotics
• Biomechanics
• Materials Science

Background:

• Accurate characterization of the human-machine interface is crucial for optimizing wearable robot performance and understanding user biomechanics.
• Existing methods for sensing this interface face challenges due to manufacturing complexity and non-linear sensor responses.
• Developing novel, reliable sensing modalities is essential for advancing human-robot interaction.

Purpose of the Study:

• To introduce and validate a novel fluidic innervation sensing modality for measuring human limb-device mechanical interaction.
• To assess the linearity and accuracy of the fluidic sensor in correlating pressure changes with applied forces.
• To demonstrate the sensor's applicability in both controlled laboratory settings and more unconstrained real-world scenarios, including integration with wearable robotic systems.

Main Methods:

• Fabrication of a 3D-printed silicone pad with embedded air channels for fluidic innervation.
• Measurement of pressure changes within the air channels in response to applied forces using off-the-shelf pressure transducers.
• Benchtop testing to establish the linear relationship between applied force and measured pressure ($R^2 = 0.998$).
• Clinical validation using a dynamometer to correlate sensor pressure with isometric knee torque ($R^2 = 0.95$ for flexion, $R^2 = 0.75$ for extension).
• Testing in unconstrained settings, including elbow flexion/extension during curls and integration into a lower-extremity exoskeleton for squat analysis.

Main Results:

• A strong linear correlation ($R^2 = 0.998$) was observed between applied force and pad pressure in benchtop tests.
• Clinical validation demonstrated high correlations between above-knee pressure and knee flexion torque ($R^2 = 0.95$), and below-knee pressure and extension torque ($R^2 = 0.75$).
• The sensor successfully captured pressure variations related to joint angles during arm curls and tracked task dynamics during exoskeleton-assisted squats, indicating robust performance in unconstrained environments.

Conclusions:

• Fluidic innervation presents a customizable and effective sensing modality for capturing human-machine mechanical interaction with high signal-to-noise ratio and temporal resolution.
• This technology offers a promising alternative for real-time sensing input to control and optimize wearable robotic systems.
• The sensor's ability to track user function during device use has significant implications for rehabilitation, assistive technologies, and human-robot interaction research.