Date of Award


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Electrical Engineering


Simon Haykin


This thesis summarizes the results of an experimental investigation into the advantages of receiver relative phase measurement, and transmitter polarization switching for the detection of co-operative polarimetric targets in ground clutter. In addition to assessing the impacts of these hardware enhancements, the investigation quantifies relative merits of linear and circular polarizations for this application. The final goal of this research is to quantify the detection performance improvements possible through both the hardware and processing enhancements described. These results will then be used to define the architecture of the next generation of polarimetric radar navigation systems. The investigation begins by reviewing the mathematical representation of polarized waves, and the ways that polarization has been exploited in past radar systems. Next, a mathematical reflectance model is developed for the proprietary trihedral twist grid reflector, and applied to predict anechoic chamber measured responses. An experimental field trip was conducted to measure this reflector and ground clutter using the National Research Council polarimetric precipitation measurement radar. The seven-dimensional radar echo measurements were described using multiple scatter plots and analyzed using the K-nearest neighbour rule pattern classification technique. The results of this analysis revealed that both transmitter polarization switching and receiver relative phase measurement should be used together, and that these two features can reduce the probability of false alarm by up to an order of magnitude without affecting probability of detection. Lastly, linear polarization was demonstrated as providing higher performance than circular polarization for the detection of tile co-operative targets in ground clutter, both using conventional amplitude detection, and also for the amplitude-correlation channel detection.

Files over 3MB may be slow to open. For best results, right-click and select "save as..."