Accuracy Analysis and Optimal Configuration for AOA Localization

This paper investigates the optimal sensor configuration problem for bearings-only target localization. A unified analytical model for localization accuracy is established, applicable to both 2D and 3D spaces. Using A-optimality as the performance metric, an eigenvalue analysis of the Fisher information matrix (FIM) reveals that the equivalent condition for optimal sensor configurations is that the eigenvalues of the FIM are equal. Under the assumption of identical measurement weights, the formation rules of optimal configurations in 2D and 3D spaces are derived. Numerical verification is conducted for a four-sensor scenario in 3D space. The sensor positions are optimized by gradient descent. Monte Carlo simulation results agree well with the theoretical Cramér–Rao lower bound (CRLB), validating the proposed analysis.