In-Sensor Tactile Fusion and Logic for Accurate Intention Recognition

Area of Science:

• Materials Science
• Computer Science
• Human-Computer Interaction

Background:

• Traditional human-machine interactions (HMIs) face challenges with redundant data and low efficiency due to separate sensing, computing, and logic modules.
• Implementing parallel-sensing modules often requires complex circuitry, leading to inefficiencies in current HMI systems.

Purpose of the Study:

• To introduce a location-and-pressure intelligent tactile sensor (LPI tactile sensor) that integrates sensing, computing, and logic for advanced HMIs.
• To develop a sensor that overcomes the limitations of traditional architectures by enabling in-sensor perception and ultrahigh-resolution action-intention recognition.

Main Methods:

• Proposed an integrated LPI tactile sensor with a layered structure, combining sensing, computing, and logic functionalities.
• Implemented in-sensor perception through feature transmission, fusion, and differentiation, minimizing data transfer.
• Achieved high-resolution sensing for both location (<400 µm) and pressure (75 Pa).

Main Results:

• The LPI tactile sensor demonstrates efficient and ultrahigh-resolution action-intention interaction, revolutionizing the von Neumann architecture.
• Achieved synchronous feature fusion and decoding for high-fidelity recognition of action and combinatorial logic intentions.
• Showcased robust privacy features and interaction intelligence through pressure enhancement, enabling real-time continuous touch action recognition.

Conclusions:

• The LPI tactile sensor offers a paradigm shift in HMI design by integrating multiple functions into a single unit.
• This integrated approach significantly simplifies data dimensionality and enhances the efficiency and accuracy of intention-driven HMIs.
• The sensor's capabilities in recognizing complex touch intentions and providing enhanced interaction open new avenues for future HMI development.