Superdirective Microphone Arrays for Real Time Traffic Measurements
MetadataShow full item record
The measurement of sound power from a car has traditionally been done by taking the car to a test site and driving it past a microphone several times. The use of microphone arrays and beamforming can make the aperture able to single out one car and follow it, taking a measurement of it, while suppressing the sound from the traffic around it.This master thesis studies two algorithms that might be used in this kind of set up: The Delay and Sum (DAS), and the Sparsity Constrained - Deconvolution Approach for the Mapping of Acoustic Sources (SC-DAMAS). This thesis tries to answer the questions: Is SC-DAMAS a better beamforming algorithm than DAS? And are they viable for a real time implementation for the measurement of cars in traffic?The SC-DAMAS algorithm achieves a far better signal to noise ratio than the DAS. In simulations with background noise levels that are 4 times higher than the source signal, the SC-DAMAS still achieved a signal to noise ratio of 10 dB. SC-DAMAS also excels in low frequency resolution: Where DAS results in a 4 m wide resolution at 400 Hz (for a 2 m long array), the SC-DAMAS manages to pinpoint the source to within one measurement field segment at 200 Hz.DAS on the other hand uses far less resources computing the results from one simulation. The SC-DAMAS used 270 times as much time concluding the simulation compared to DAS. In light of this result it is questionable whether it is possible to implement the SC-DAMAS in a way that lets it measure the sound power levels from a car in real time.