Experimental validation of a tactile sensor model for a robotic hand

Area of Science:

• Robotics
• Biomimetic Engineering
• Sensor Technology

Background:

• Human hands possess sophisticated tactile sensing capabilities through mechanoreceptors.
• Replicating these natural sensors in robotic systems is crucial for advanced manipulation.
• Existing robotic tactile sensors often lack the nuanced response of biological systems.

Purpose of the Study:

• To design and evaluate a novel tactile sensor for robotic hands.
• To emulate both slow-adapting and fast-adapting mechanoreceptors found in human glabrous skin.
• To enable accurate detection of tri-axial force components.

Main Methods:

• Utilized strain gauges to replicate slow-adapting mechanoreceptor function.
• Employed polyvinylidene fluoride (PVDF) film for fast-adapting mechanoreceptor simulation.
• Integrated four strain gauges and one PVDF film beneath a force-localizing protrusion.
• Oriented strain gauges for tri-axial force component identification.
• Applied multiple linear regression with interaction terms for force prediction.

Main Results:

• The sensor unit successfully integrated strain gauges and PVDF film.
• The protrusion effectively localized force to the sensor's receptive field.
• The developed regression model accurately predicted tri-axial force components.
• Achieved mean percentage errors below 15% for each force component.

Conclusions:

• The biomimetic tactile sensor effectively replicates human glabrous skin mechanoreceptor functions.
• The sensor demonstrates high accuracy in detecting multi-axial forces for robotic applications.
• This technology advances the development of dexterous and sensitive robotic hands.