The design of irregular subarrayed planar phased arrays composed of modular tiles is a research area of increasing interest. In this article, a heuristic iterative convex relaxation programming (H-ICRP) framework is proposed. By formulating the tiling of irregular subarray as a binary programming problem, this framework is suitable for arbitrary array aperture and subarray structure. To achieve exact aperture cover, the nonconvex binary constraint is relaxed to a convex approximation and solved through multiple iterations. Furthermore, for better performance of the scanning radiation pattern, the multitask Bayesian compressive sensing (MT-BCS) solver is adopted to find a heuristic initial iteration point instead of randomizing, such that the robustness and effectiveness can be improved remarkably. Good performance and potentialities of the proposed method are verified by comparing with existing approaches in various numerical examples.

Subarray design technique is widely employed in the design of planar phased arrays. However, it's hard to obtain full coverage and good radiation performance with the given shape of subarrays and array aperture. In this letter, the subarray design problem is formulated as an iterative convex relaxation model and solved through some steps subtly. Besides, the scanning sidelobe level (SSL) is adopted to retrieve the optimal configuration among the found solutions. Numerical experiments are carried out to assess the performance of the proposed method as well as the inherent superiority in adaptive beamforming.

For a dwell time limited radar network, the achievable tracking performance can be further enhanced by adaptively tuning in the dwell time allocation strategy to accommodate the future operation scenario. In this paper, considering this dynamical resource allocation problem, a cone programming-based dwell time allocation scheme is proposed for achieving the optimal tracking performance. The basic mechanism is to exploit the feedback of its operating environment based on the interrogation of received echoes for informing the decision of future optimal dwell time allocation. To achieve this purpose, the predicted conditional Cramer-Rao lower bound (PC-CRLB) of target state estimates is developed for evaluating the tracking performance of each candidate allocation strategy conditioned on the cognitive knowledge firstly. And, its trace is chosen as a scalar allocation strategy evaluation metric. On this basis, the closed-form expression of the scalar evaluation metric is derived, which enables us to transform the dwell time allocation problem into the second-order cone program (SOCP). Extensive simulations demonstrate the efficiency and superiority of the proposed allocation scheme.

Distributed radars have the potential to combine coherently for achieving a high signal-to-noise ratio (SNR) while maintaining a moderate antenna size. The key to coherently combining multiple radars is obtaining accurate coherent parameters (CPs), which are used to adjust the transmitting/receiving time and phase of each radar. One approach for CP estimation is to transmit orthogonal waveforms. However, ideally, orthogonal waveforms occupying the same frequency band may not be found in practice. Cross-correlation energy leakage exists between non-orthogonal waveforms, which seriously impairs the accurate acquisition of CPs. To solve this problem, we propose a clean signal reconstruction approach for CP estimation. This approach reconstructs clean echoes by gradually stripping out the cross-correlation energy leakage with a reconstruction-elimination-reconstruction framework. And CPs are obtained from these reconstructed clean echoes. Since the majority of cross-correction energy leakages are eliminated, enhanced CP estimation performance can be achieved. Verified simulations are designed for a dual radar scenario. Results show that the proposed approach significantly improves the performance of CP estimation while reducing the SNR requirement for coherently combining multiple radars.

Frequency diverse array (FDA) can offer new degrees-of-freedom, but the target localization of basic FDA is fundamentally limited by the range-angle-coupled and rangeperiodic peaks of beampattern. In this paper, we propose a localization system with the combination of subarray-based FDA and full-band FDA (FB-FDA), as the transmitter and the receiver respectively. The transmit and receive beampatterns are both decoupled in range and angle domains. Target rangeangle localization is achieved by two steps. Firstly, using the outputs of subarrays, we solve the range ambiguity and obtain a coarse location. Secondly, the monopulse processing of rangeangle localization is developed for the proposed system by taking the coarse location as the receiving beam pointing. The localization performance is also analyzed in Cramér-Rao lower bound (CRLB), computational complexity and output signal to interference-plus-noise ratio (SINR). Several simulation results demonstrate that unambiguous target range-angle localization can be achieved with a low computational cost by the proposed method.

In this letter, a general framework of cognitive resource allocation for target tracking in radar networks is proposed. Firstly, an allocation strategy evaluation metric is established in this framework, i.e., the predicted conditional Cramer-Rao lower bound (PC-CRLB), for evaluating the future response of each candidate allocation strategy. Then, the optimal resource allocation strategy can be found by solving a constrained optimization problem. To specify the framework implementation details, a dwell time allocation problem is considered. For this given problem, the analytical expressions of PC-CRLB and its derived scalar strategy evaluation metric are derived, enabling us to convert the dwell time allocation problem to a second-order cone program (SOCP). Numerical results demonstrate the effectiveness of the proposed framework.