Search
Search
The recognition of sound source characteristics such as the volume, the
direction and the distance between a sound source and sensors is an
important technology for unmanned systems like autonomous vehicles,
robot systems and AI speakers. The conventional sound source recognition
algorithm by the crossing shape recognizes the direction of a sound source
but this algorithm can’t recognize the distance between a sound source
and sensors.
ASSR Algorithm for AI Sound Sensor System
The ASSR algorithm uses 6 pairs of interaural time difference which is
calculated based on the quadrature right-angle acoustic sensing by the
rectangular shape. The ASSR algorithm can solve the existing problems
of the crossing shape such as occurring of blind spots, and calculate the
distance between a sound source and sensors. Experiments show that the
proposed ASSR algorithm can recognize the volume, the direction and the distance between a sound source and sensors without blind spots.
According to the test, the ASSR algorithm computed the volume, the
direction and the distance between a sound source and sensors, and
recognized the sound source characteristics. Derived the performance
model of the proposed algorithm, it showed that the recognition rate
of direction reached up to 96% ~ 100% depending on the direction
without any blind spot, and also the detection rate became 97% at 20 dB
signal to noise ratio. The error rate of the computing distance between
the sound source and a sensor met within +/- 5% error rate. The ASSR
algorithm could also compute the distance between a sound source and
sensors even when the sound source is located at a long distance without
a blind spot. In conclusion, it confirms that the ASSR algorithm can introduce an auditory function into the unmanned system.
