A Processing Technique for OFDM-Modulated Wideband Radar Signals

Abstract

The orthogonal frequency division multiplexing (OFDM) is a multicarrier spread-spectrum technique which finds wide-spread use in communications. The OFDM pulse compression method that utilizes an OFDM communication signal for radar tasks has been developed and reported in this dissertation. Using the ambiguity function tool, the feasibility of the OFDM pulse compression method was demonstrated from a performance perspective. The two fundamental components of the OFDM communication signal, namely the cyclic prefix guard interval and the random message content, are incorporated in the OFDM radar signal and the signal processing technique is developed to make use of these features rather than avoid them. The structure of the multicarrier signal and the unique outcome of the novel processing technique offer a new solution to the ambiguity in the Doppler measurements. The Doppler effect, which is considered as a cause of pulse compression loss, is compensated for in the frequency domain, and the additional information on the Doppler effect sue to the multicarrier structure helps solve the Doppler ambiguity occurring in the coherent integration stage. Two peakpower limiting techniques were considered for peak-power reduction; one method modulates the OFDM carriers by Golay complementary codes while the other method applies pre - coding by the DFT matrix to obtain the single-carrier OFDM (SC-OFDM). The assessment of these PAPR-control methods for radar applications generated novel results, and answers the valid concerns regarding the linearity of the power amplifiers. With higher radial velocity, the point target does not occupy a single range bin during the extent of the radar signal but migrates from one range bin to the next. The range migration occurs first at the level of coherent Doppler integration, where the OFDM pulse compression can still assume the Doppler effect to be a frequency shift. At higher velocities, the actual Doppler effect that is the scaling on the signal cannot be accurately modeled as a frequency shift anymore. Compensation techniques for both effects were developed in this thesis and verified through simulations.