Designing a simple hearing system for a humanoid robot
The shortest way to avoid humans being busy with their daily routine is to develop humanoid robots with sensitive features. However, it is complicated when considering all functionality together. Therefore, separate studies of different body parts are taken into account. Meanwhile, designing a hearing system for a humanoid robot is interesting and it gives a sense of vitality to the robot. Previously, there were many approaches to design robot-hearing systems. This blog post describes a simple method that can be used for a robot-hearing system.
Theoretical background
Sound Source Localization (SSL) is a technique that uses to find the location of a sound source. To exactly find the location of a sound source, three minimum microphones are needed. But, to find the angle of arrival of an acoustic signal, two microphones are enough. When someone speaks to us, we look in the direction that the sound came from. That capability can be added to a humanoid robot using the Angle of Arrival (AOA) technique. That can be done easily with the help of the following three assumptions.
Assumptions:
Sound waves detected by two microphones are parallel to each other.
The change in the velocity of sound in the air due to variations in pressure, temperature, and humidity is negligible.
The effect of reverberation is negligible.
According to Figure 1, signal A takes Δt time more than signal B to reach the microphone. If we know that Δt time we can find the θ angle using the equation shown below.
θ = arcos (Δt * v/d)
Where,
θ = Angle of arrival
d = Distance between two microphones
Δt = Time difference of arrival
V = Velocity of Sound in air
But, how we can measure this very small Δt time?
Generally, this Δt time is referred to as the Time Difference of Arrival (TDOA). The method called cross-correlation can be used to determine TDOA. The cross-correlation measures the similarity between two signals. So, the peak point of the cross-correlation result gives the TDOA.
Now, the θ angle can be found using the equation I mentioned above.
Thanks for reading. If you enjoyed this article, please hit the clap icon.:)
References
J.T.A. Raghesh, G.V. Rao, K. Saravanan, “Acoustic Source Localization and Navigation Robot,” in International Journal of Engineering and Advanced Technology (IJEAT), 2019.
S. Paulose, E. Sebastian, Dr. B. Paul, “Acoustic Source Localization,” in International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, 2013.
S. Ramnath, G. Schuller, “Robust Speech Direction Detection for Low-Cost Robotics Applications,” in Multi-Learn, Kos island, Greece, 2017.
