Vehicular Classification Based on Vibration Caused by Uncontrolled Traffic Using Fibre Optic Sensor

Abstract

Fibre optic vibration sensor (FOVS) converts vibration signal to light signal. Due to its prominent features, distributed fibre optic vibration sensor is preferred to conventional methods. Interest in FOVS has greatly increased over the years in structural health monitoring and vehicular traffic, hence the need to embark on this study. Distributed FOS is employed to measure the frequency of vibration caused by uncontrolled vehicular movement at Oluku By-Pass Bridge. ?-OTDR is used to obtain millisecond snapshots of the stochastic signal arising thereof. A video shot was also recorded to match the exact timing of each excitation.

Traces obtained show several frequency peaks at various corresponding backscatter level for high and low vehicular traffic. Sampled data analyzed with Fiberizer Cloud software indicates high attenuation contribute to low total loss and low attenuation lead to high total loss. Spectral analyses of the data at low and high traffic for the stochastic signal. The corresponding frequency peaks were calculated and the results used to classify vehicles at low and high speed. The FWHM obtained with double Gaussian model for differential trace shows that high traffic gives sharp peaks with standard deviation less than 0.6 and above 0.6 for low traffic. The analyses identified peaks above 4.0x10-3 for traces with trucks and cars at high speed, while peaks less than 4.0x10-3 were obtained for traces with cars.?