An automatic siren detection algorithm using Fourier Decomposition Method and MFCC

In this work, an automatic ambulance detector has been proposed using features extracted from the siren of the ambulance. The smart ambulance detector can reduce the time required by an ambulance to reach their destination in emergency situations. In the present scenario, the cars are designed to provide more and more luxury to the driver and insulation from the outside noise. Along with the traffic noise, the siren sounds also get muffed, leaving the drivers clueless to the emergency vehicles in their vicinity. In these cases, the proposed system can be used to alert the driver about the approaching emergency vehicle. The proposed system uses audio sensors to record the siren sound and pre-process the acquired signal using a bandpass filter. In this work, two sets of features are computed, the first set consists of Mel-frequency cepstrum coefficients of the filtered signal. For the second set, the signal is decomposed in frequency domain using Fourier decomposition method. Statistical features such as kurtosis, energy, and variance are computed from each of the sub-bands. Relevant features are selected based on the Kruskal-Wallis test. The selected feature set is then used to train a machine learning model to identify siren sounds from the background traffic noise. In this work, we have compared the performance of various machine learning algorithms like kNN, SVM and ensemble bagged trees to select the best model. The dataset used in this work includes signals from two publicly available datasets with ambulance siren audio files and traffic sound files, and also audio data collected from different sources on the internet. We have also recorded the siren sounds in city traffic conditions.