Modeling An Instrumented (Tractor-semitrailer Combination) Vehicle: Using System Identification Methods

Abstract

In this thesis, properties of the input-output relationships of a tractor-semitrailer combination vehicle were analyzed. Predictive models of the system were identified using only input-output data collected separately but for the same portion of the road. A vehicle equipped with an inertial profiling system and a strain gauge instrumented truck were used to collect the input and output data, respectively. Input-output combinations such as single-input-single-output (SISO), multiple-input-single-output (MISO), and multiple-input-multiple-output (MIMO) representing the vehicle models were investigated using the polynomial and state-space black-box modeling (i.e., system identification) techniques. Comparisons among measured and k-steps ahead predicted outputs and residual analysis were used for model validation. These preliminary studies suggest that such a system, for prediction purposes, can be modeled using linear dynamic models based on only input-output data. Researchers and practitioners can use the study results to develop applications related to pavement-vehicle interaction.