Approximate Maximum Likelihood Estimator of Multi-target Motion Parameters for Orthogonal Frequency Division Multiplexing Radar

To detect multiple targets on a phase-coded Orthogonal Frequency Division Multiplexing (OFDM) radar, this paper proposes a parameter estimation method based on channel separation and maximum likelihood principle. First, the multi-channel signals were separated because of the orthogonality of the OFDM system, and then the separated signals were correlated with the phase-coded reference signal in the fast time domain to acquire a 1-D range profile from each channel. Subsequently, Keystone transform was used to correct the effect of the coupling between Doppler shift of the subcarrier and the slow time domain. Simultaneously, coherent accumulation was conducted jointly in the slow time and subcarrier domains to obtain a 2-D range-Doppler spectrum. Using CLEAN technique, the peaks of this spectrum were examined to obtain the range cells and velocities of each target. Using these parameters as the initial values, the likelihood function was maximized using Newton’s iterative algorithm to yield an approximate maximum likelihood estimator of the motion parameters. Simulation results demonstrate that the proposed algorithm outperforms the traditional Keystone-based estimation algorithm both in computational complexity and parameter estimation accuracy. The algorithm improved the input SNR by approximately 4 dB under the same root mean square error, and the mean square error approached the Cramer-Rao lower bound.