A Sequence-Filter Joint Optimization

Joint Design of Radar Transmit Waveform and Mismatched Filter with Low Sidelobes

“…5. It can be seen from the figure that PSL of the IMDA (−49.27dB) is significantly lower than that obtained by the algorithms proposed in [9] (−34.18dB), [10] (−41.37dB), and [11] (−40.88dB). A lower PSL can be obtained by choosing a good initialization, such as Golomb sequences shown in Section IV-A (−58.07dB).…”

“…The histograms indicate that the PSLs corresponding to the transmit waveform and mismatched filter designed by the IMDA initialized for different initial CAN sequences are different. However, compared with the methods of [10] and [11], the IMDA produces lower PSLs.…”

“…Then r l requires about N flops. Therefore, the computational complexity of updatingâ [it] is O(|Ω|(M + 2p log 2 (2p))), which is the same as that for the methods in [10] and [11]. Moreover, the IMDA can be accelerated by using SQUAREM [16] which requires about |Ω|(M + 2p log 2 (2p)S) flops per MM iteration, wherē S is a mean iteration number of SQUAREM used.…”

“…The parameters of the IMDA are p = 10, δ o = δ i = 10 −5 , and other parameters are set the same as those in Section IV-A. For comparison, we also run the methods proposed in [10] (with optimally selected parameters α = 0, β = 0.3, m = 20, and µ = 2) and [11] (with optimally selected parameters ρ = 20, µ = 5 × 10 −5 , and pattern g is of −70dB PSL) with the same 100 initial CAN sequences. The histograms of PSLs for different methods tested are plotted in Figs.…”

“…Joint design of transmit waveform and mismatched filter is an alternative way to reduce sidelobes. In [10] and [11], gradient descent-based methods for joint transmit waveform and receive mismatched filter design have been developed to minimize the peak sidelobe level (PSL).…”