A Highly-Sensitive Omnidirectional Acoustic Sensor for Enhanced Human-Machine Interaction.

Area of Science:

• Robotics and Human-Machine Interaction
• Materials Science and Sensor Technology
• Acoustics and Signal Processing

Background:

• Acoustic sensor-based human-machine interaction (HMI) is vital for intelligent robots but faces challenges in identifying and tracking omnidirectional sound sources, especially in noisy environments.
• Existing systems struggle with accuracy and robustness when dealing with complex acoustic scenes.

Purpose of the Study:

• To develop a self-powered triboelectric stereo acoustic sensor (SAS) with enhanced omnidirectional sound recognition and tracking capabilities.
• To address the limitations of current acoustic HMI systems in noisy and complex environments.

Main Methods:

• A 3D-structured triboelectric stereo acoustic sensor (SAS) was designed and fabricated.
• The sensor utilizes a porous vibrating film with specific material properties (high electron affinity, low Young's modulus) for high sensitivity and wide frequency response.
• Deep learning algorithms were employed for audio signal identification and tracking.

Main Results:

• The developed SAS demonstrated high sensitivity (3172.9 mV_pp Pa^-1) and a wide frequency response range (100-20 000 Hz).
• Achieved an average deep learning accuracy of 98% in precisely identifying desired audio signals, even in noisy environments.
• Successfully demonstrated simultaneous recognition of multiple individuals and driving commands in complex scenarios like conference systems and self-driving vehicles.

Conclusions:

• The self-powered triboelectric SAS offers a significant advancement for voice-based HMI systems.
• The sensor's omnidirectional recognition and noise resilience pave the way for more natural and efficient human-robot communication.
• This technology holds promise for applications requiring robust audio-based interaction in challenging acoustic conditions.