Machine Learning-Enhanced Modular Ionic Skin for Broad-Spectrum Multimodal Discriminability in Bidirectional
Qianqian Yang1, Bingqiao Li2, Mengke Wang1
1State Key Laboratory of Fluid Power and Mechatronic Systems, School of Mechanical Engineering, Zhejiang University, Hangzhou, 310000, China.
Advanced Materials (Deerfield Beach, Fla.)
|July 21, 2025
Summary
A new machine learning-enhanced ionic skin offers broad-spectrum multimodal sensing by optimizing sensors and algorithms. This advanced tactile perception system improves human-machine interactions with enhanced temperature and pressure discrimination.
Area of Science:
- Materials Science
- Robotics
- Artificial Intelligence
Background:
- Current multimodal tactile perception systems face challenges with limited sensing ranges and decoupling strategies.
- Effective multimodal sensing is crucial for advanced human-machine interactions.
Purpose of the Study:
- To develop a machine learning-enhanced modular ionic skin (MIS) for broad-spectrum multimodal discriminability.
- To overcome limitations in sensing range and decoupling for tactile perception systems.
Main Methods:
- Developed a synergistic sensor-algorithm optimization strategy for the MIS.
- Engineered ionic conductors through process-controlled hard-segment modulation in ionic gel.
- Proposed and trained a data-driven decoupling model using a multi-stimuli dataset.
Main Results:
- Achieved enhanced sensing properties: minimum temperature coefficient of -4.00% °C⁻¹, linear gauge factor of 2.95, and maximum pressure sensitivity of 80.5 kPa⁻¹.
- Demonstrated maximum decoupling ranges for temperature and pressure with prediction errors as low as 7.0%.
- Maintained reliable strain detection under temperature interference.
Conclusions:
- The MIS system exhibits effective multimodal sensing capabilities.
- The developed system shows potential for applications in bidirectional human-robot interaction, including wearable hand kits and robotic grippers.
