Vehicle Classification for Facility Monitoring, Los Alamos
National Labs, June 2007
The goal of the project was to build a low
energy, reliable classification system to be deployed at remote access points that
can distinguish vehicles into 3 main classes: personal, light weight vehicles
such as cars, medium load carrying trucks and heavily loaded trucks. For this
project I have designed and implemented a spectral characteristics based
classifier using a Stargate / Mica2 pair for
processing and communication and acoustic and seismic sensors. The seismic
sensor connected to a mica2 with low sampling frequency act as a trip wire
which enables the acoustic sensing based classification system on the Stargate. The FLDV based classifier implemented on the Stargate performed with 100% classification accuracy in
over 20 runs with each vehicle class. (Detailed
slides)
Publications
1.
J. Frigo, V. Kulathumani, S. Brenna, E. Rosten,
E. Raby, Sensor Network Based
Vehicle Classification and License Plate identification System, INSS 2009