Embodied AI Glossary中文

Microphone Array

麦克风阵列Advanced

Multiple microphones arranged in a fixed geometry working together to pick up sound directionally and judge where it comes from.

A microphone array is a set of microphones arranged in a known geometry — a line, a ring, and so on — that capture and jointly process signals together. Sound reaches each microphone at a slightly different time, which can be used to estimate the direction a sound came from (sound source localization, also called direction-of-arrival or DOA estimation); delaying and weighting the different channels before combining them can boost sound from one particular direction while suppressing noise from others — this is called beamforming. A single microphone can do neither of these. Microphone arrays are widely used in phones, smart speakers, hearing aids, and as the front end for speech recognition. For robots, a microphone array lets it pick out a command clearly in a noisy environment, judge where a speaker is and turn to face them, and it is also the hardware foundation for “listen to find the target” research like audio-visual navigation. Many humanoid robots treat it as standard equipment — for instance, Unitree’s G1 spec sheet lists a 4-microphone array.

ExampleA user calls the robot’s name from across the living room; the robot estimates the sound’s direction with its microphone array, turns to face the user, then uses beamforming to boost speech from that direction before sending it to speech recognition.

Also called
Sound Source Localization, Beamforming
Related
Automatic Speech Recognition · Audio-Visual Navigation · Multimodal Perception · Human-Robot Interaction · Multi-Sensor Fusion · Contact Microphone
Sources
Wikipedia: Microphone array
Unitree G1 产品页(规格表:4 Microphone Array) (Chinese)
As of
2026-09

See it in the full glossary →