Full text loading...
Sketch of the first experiment. A source sends a broadband pulse in an open medium. It is sensed by a fixed and nearly omnidirectional receiver . Two diameter empty cylinders are placed nearby. The source is moved around the setup (2000 different positions), along a circle of radius .
Individual contribution of each of the 2000 sources to the total Correlation , on a log scale. As noticed by Snieder (see Ref. 15) based on a stationary phase approach, sources efficiently contributing to the Green’s function are in the receiver-reflector direction. The wave fronts labeled [(a)–(b)–(c)–(d)] correspond to the direct (a) and secondary [(b)–(c)–(d)] arrivals in the retrieved Green’s function (Fig. 3).
(a) Autocorrelation averaged over 2000 sources, showing a strong peak at time . [(b)–(c)–(d)] Zoom into the late times. Thin dotted line: averaged autocorrelation ; thick line: reference impulse response .
Setup of the second experiment and typical wave forms. The source shoots a pulse, the wave is sensed in ; the same transmit-receive procedure is repeated for each of the 118 available sources. The receiver’s directivity pattern is represented (thin line). At each position it is rotated to capture the field from various angles and compensate for its lack of omnidirectionality.
118 source-averaged correlations for 12 distances . The direct and reflected signals are superimposed. Dotted lines represent the arrival times for the direct and reflected paths. The scatterer is a plane interface positioned away from the sensor array.
Article metrics loading...