EFFICIENT SENSING OF GROUND-BORNE VIBRATIONS INDUCED BY PILE DRIVING USING COMPRESSIVE SAMPLING

In this paper, we propose and discuss the applicability of the compressivesensing (CS) for the measurement and analysis of the ground-borne vibrations inducedby pile driving. With this motivation, we consider two types of the most common piledriving techniques; i) impact pile driving and ii) vibratory pile driving. We show thatground-borne vibration parameters like longitudinal velocity (LV), transverse velocity(TV), vertical velocity (VV), force and inertia can be efficiently sensed by the CS forboth of the two driving types mentioned above, since majority of these parameter's timeseries are sparse either in time or frequency domain. We show that time series of theseparameters can be reconstructed using a far fewer number of samples than the classicalShannon's theorem states. Thus, when the CS technique is utilized, measurement timesand hardware storage requirements of vibration sensing systems can be significantlyreduced. Additionally, when the CS is used, such parameters can be exactly reconstructedeven in the presence of missing data. CS based proposed in this paper can also beextended to measure and analyze other types of vibration data in soil dynamics andmore generally to other vibration types encountered in engineering.