Feasibility of implementing data fusion techniques in non-destructive testing of pavements

Robert Reuben Williams, University of Texas at El Paso


There are several non-destructive testing devices and methods that are currently being utilized to evaluate the properties of pavement layers, namely modulus and thickness. The decision of which data to put more credence in arises when there is more than one option. One possible solution to this situation is to use data fusion to integrate the data in a synergistic way, such that the resulting output will take into consideration all available data and create a more reliable perception of the pavement system. The primary testing devices currently in use are the falling weight deflectometer, the seismic pavement analyzer, the portable seismic pavement analyzer, and ground penetrating radar, all of which provide raw data that can be used to obtain thickness or modulus information. As each device has unique strengths and weaknesses, logic only dictates to utilize the strengths of each method while minimizing the weaknesses. Results do not always coincide and thus decisions on either to combine values or decide on one value need to be made. Data fusion is one potential solution to this dilemma. Data fusion allows for the combining and filtering of information to obtain a composite value or a basis for decision. There are many forms of data fusion, yet not all methods of data fusion are applicable to the combining of NDT pavement data. Data fusion has yet to be applied to the non-destructive testing of pavements. This thesis presents several data fusion alternatives which will show the potential for using data fusion for the synergistic analysis of data from multiple non-destructive testing devices.

Recommended Citation

Williams, Robert Reuben, "Feasibility of implementing data fusion techniques in non-destructive testing of pavements" (2003). ETD Collection for University of Texas, El Paso. AAIEP10617.