Bio-Acoustic Tracking and Localization using heterogeneous, scalable microphone arrays
Erik Verreycken, Ralph Simon, Brandt Quirk-Royal, Daems, Walter, Jesse R. Barber, Jan Steckel · Zenodo (CERN European Organization for Nuclear Research) · 2021
Contents of repository * Figure 1e Data file: figure_1e.mat Variables: - bk_fs: Sample rate used for recording with the Bruel and Kjaer (BnK) - bk_y: Single channel audio data of BnK recording - mems_y_resampled: 3 channel recording made with the mems microphones. Resampled and cut to match the BnK recording. Code file: figure_1e.m Matlab code that provides an example on how to read the data and display the graph seen in figure 1e. * Figure 3 (b, c, d, e, f, g) Data file: figure_3.mat Matlab data file containing the following variables: - data: 64 channels (8 devices * 8 microphones) of acoustic data from a single bat-prey interaction. data contains both the matched filter output data and the envelope detection data. - peakStart and peakStop: The sample indexes that indicate the bat call used in this example - fs: sample frequency of data used here - dataPeakEnvThresh: Calculated thresholds for peak detection - dataPeakTimes: Sample indexes that a peak or bat call was detected - AllPeakThresh: Calculated threshold for the combined peak data - call1, call2: 2 selected bat calls for visualization in a spectrogram - flightPath: Probabilistic data of bat positions during a single approach - micPos: 3D positions of the microphones in this experiment Code file: figure_3.m Matlab code that provides an example on how to read the data and display the graph seen in figure 3 subfigures b,c,d,e,f,g. * Figure 3hi and figure S2 Data file: figure_4.mat: Matlab data file containing the following variables, all variables here have 1787 rows, corresponding to the 1787 calls (821+966) in the manuscript: - batID: List of matching bat ID's for each call. - callCnt: Index of the call (resets per approach) - flightID: Unique ID per flight (up to 121, corresponding to the manuscript) - FOA: Center position of emission (in meter) - focusedness: focusedness value for each call - scorpPos: position of the scorpion on the board (x, y in meter) - surfaceType: surface type for this flight (1=rough, 2=smooth) Code file: figure_3hi_S2.m Matlab code that provides an example on how to read the data and recreate subfigures from the manuscript * Figure 4 and figure S3 Data file: figure_4.mat: Matlab data file containing the following variables: - birdCapture: Contains array of 47 bird captures - capture time - call info - DOA location - TDOA location - micPos: 3D locations of all 100 microphones - call_chaffinch: A single chaffinch song - call_dunnock: A single dunnock song - doa_sphere_pts: Uniformly spaced samples on a 3D hemisphere. This sphere indicates the directions that are sampled to find the most likely point of origin. By using a more dense sphere the algorithm can be made more precise at the cost of being computationally more intensive. - doa_L: Likelihood of direction being the sound direction for all directions from doa_sphere_pts and for all devices. - device_locations: 6D position of all the devices. x, y, z in meter, Roll, pitch and yaw in radians. - doa_rotations: Relative rotation to map the doa_sphere_pts over each device. Code file: figure_4.m Matlab code that provides an example on how to read the data and recreate subfigures from the manuscript bg3.png: Image used as background for the DoA and TDoA figures. * Figure S1 Data file: figure_S1.mat: Matlab data file containing the following variables: - scans: 22 scan patterns from a single approach. Code file: figure_S1.m Matlab code that provides an example on how to read the data and recreate figure from the manuscript * Figure S5 Data file: figure_S5.mat: Matlab data file containing the following variables: - birdSong: A single bird song - fsBird: Sample rate of the bird song - batCall: A single bat call - fsBat: sample rate of the bat call Code file: figure_S5.m Matlab code that provides an example on how to read the data and recreate figure from the manuscript * Figure S6 abcde Data file: figure_S6.mat: Matlab data file containing the following variables: - dataPsd: psd of data for 180 degree rotation in 1 degree intervals, expressed in dB - fmSweep input signal of fm sweep from 80 to 20kHz in 10ms - allFOA_dist_gt: PDF of simulated FOA information comparing the distances between the ground truth FOA position and the offset introduced by microphone directionality losses - allFOA_dist_rand: PDF of FOA error introduced by random amplitudes introduced by random angles. - allFOA_dist_angle: PDF of FOA error introduced based on realistic angular variations. Code file: figure_S6.m Matlab code that provides an example on how to read the data and recreate figure from the manuscript * Figure S7 abc Data file: figure_S7.mat Variables: - fs: Sample rate of the recorded audio data - y: 6 channel audio data from experiment - sig: Signal as played by the speaker Code file: Matlab code that provides an example on how to read the data and display the graph seen in figures S7 a.b and c.