Spherical-Wave Steering Vector Analysis

• Spherical-wave steering vectors are defined as the array response vectors expressed via vector spherical wave functions, capturing near-field curvature and complex propagation effects.
• They provide an exact and numerically stable framework for modeling electromagnetic fields, encompassing wideband, polarization-dependent, and mutual coupling characteristics.
• The approach underpins advanced applications in beamforming, near-field localization, and spatial sampling optimization, influencing modern antenna performance analysis.

A spherical-wave steering vector is the array response vector for a given source or field, expressed via an expansion in vector spherical wave functions (VSWFs). This formalism generalizes the classical plane-wave steering vector: in the near field or under complex propagation and scattering, the field incident on an array cannot be modeled as locally planar; instead, the full spherical (or, in the most general setting, arbitrary) Helmholtz solutions must be used. The VSWF basis enables an exact and numerically stable description of the array’s response that accommodates wideband, polarization-dependent, and mutual-coupling effects, as well as arbitrary incident field geometries and boundary conditions. Spherical-wave steering vectors are foundational in advanced electromagnetic array modeling, spatial signal processing, antenna performance analysis, near-field localization, and rigorous implementations of the Cramér–Rao bound in array parameter estimation.

1. Mathematical Structure of the Spherical-Wave Steering Vector

The spherical-wave steering vector is constructed by decomposing any incident source-free electromagnetic field into a sum over VSWFs, specifically the $\mathbf{M}_{nm}(k\mathbf{r})$ and $\mathbf{N}_{nm}(k\mathbf{r})$ families. In canonical form, any field $\mathbf{E}_{\text{inc}}(\mathbf{r})$ can be expanded as

$$\mathbf{E}_{\text{inc}}(\mathbf{r}) = \sum_{n=1}^\infty \sum_{m=-n}^{n} \left[ a_{nm} \,\mathbf{M}_{nm}(k\mathbf{r}) + b_{nm} \,\mathbf{N}_{nm}(k\mathbf{r}) \right]$$

where $(a_{nm}, b_{nm})$ are the beam-shape coefficients (BSCs), which collectively form the spherical-wave steering vector or "phase-mode" vector (Moreira et al., 2010). Outgoing and incoming (regular) VSWFs are distinguished based on the requirement of finiteness at the origin and radiation conditions at infinity. The normalizations and explicit formulae for VSWFs are given in closed form in terms of spherical Bessel and Hankel functions and scalar spherical harmonics.

2. Physical Modeling and Key Features

VSWF-based steering vectors capture not only the geometrical path differences due to source–array configuration, but also encompass:

• Frequency, direction, and polarization dependence of each element’s reception pattern.
• Effects of mutual coupling, via the reception coefficient matrix $R_{j}^{(l)}(\omega)$ and port-coupling $\Gamma$ blocks of the generalized scattering matrix (GSM).
• Array placement and mechanical orientation, as these are encoded into the magnitude and phase of the reception coefficients.

The array’s noise-free data model, omitting amplitude and pulse-shape terms, is summarized as

$$\mathbf{a}(k,\theta_0,\phi_0,\mathbf{P}) = \left( \mathbf{1}_L \otimes e^{-i(k r_0 + \text{angle})} \right) \odot \left[ R_{j}^{(l)}(k)\,(-1)^{m[j]}\,K_{\hat{j}[j]}(\theta_0,\phi_0) \right] \mathbf{P}$$

where $K_{smn}(\theta, \phi)$ encodes the far-field pattern for each modal index and $\mathbf{P}$ denotes the polarization vector. This construction is fully general and appears in (Lafer et al., 19 Sep 2025).

3. Exact and Numerically Stable Computation

The derivation of the BSCs $\mathbf{N}_{nm}(k\mathbf{r})$0 required for the spherical-wave steering vector, historically complicated by unwanted radial dependencies, can be made radially independent by using the angular spectrum of the field in $\mathbf{N}_{nm}(k\mathbf{r})$1-space and the orthogonality properties of the VSWFs (Moreira et al., 2010). The exact expressions are:

$\mathbf{N}_{nm}(k\mathbf{r})$2

$\mathbf{N}_{nm}(k\mathbf{r})$3

This approach is numerically stable and efficient, allowing closed-form or algorithmic computation for all physically relevant beams: plane waves, Bessel beams, waveguide and cavity modes, etc. The spherical-wave steering vector is then suitable for direct use in beamforming, near-field sensing, scattering, and Mie/T-matrix calculations.

4. Near-Field and Spherical-Wave Effects on Finite Arrays

For arrays of finite extent and for sources in the near field, the phase profile on the array aperture must account for spherical curvature. This leads to a spatial chirp structure in the steering vector:

$\mathbf{N}_{nm}(k\mathbf{r})$4

where $\mathbf{N}_{nm}(k\mathbf{r})$5 is the unit vector toward the source and $\mathbf{N}_{nm}(k\mathbf{r})$6 are element positions. The first exponential is a linear (plane-wave) phase; the second is a quadratic (spherical) correction, explicitly demonstrating the spatial chirp (Monnoyer et al., 8 May 2025). For canonical structures such as Uniform Linear Arrays (ULA) and Circular Arrays (CA), this leads to straightforward expressions capturing main-lobe curvature and out-of-plane coupling.

5. Mutual Coupling, Distortion, and Calibration

Mutual coupling and nearby-object-induced distortion are fully captured by the reception coefficient matrix $\mathbf{N}_{nm}(k\mathbf{r})$7, which must be determined for each array via full-wave simulation or measurement. Port coupling is described via $\mathbf{N}_{nm}(k\mathbf{r})$8 in the GSM; mechanical changes rotate and phase-shift the $\mathbf{N}_{nm}(k\mathbf{r})$9. This ensures that the steering vector reflects the true frequency, polarization, and orientation dependence of each individual element—including all pattern and coupling effects---as emphasized in (Lafer et al., 19 Sep 2025).

6. Practical Applications: Array Optimization and Sensing

The VSWF-based spherical-wave steering vector underpins rigorous performance analysis of antenna arrays in localization, wireless sensing, and radar. Cramér–Rao lower bounds for delay and Angle-of-Arrival (AoA), accounting for channel noise and non-ideal array physics, require the VSWF steering vector as a core input. Optimization of sparse arrays for maximal localization precision (under CRLB) utilizes this formalism for realistic, compact, and performance-optimized array design, as in the context of Crossed Exponentially Tapered Slot (XETS) antennas (Lafer et al., 19 Sep 2025). The radially independent, closed-form computation of steering vectors also enables efficient simulation and real-time parameter retrieval (Moreira et al., 2010).

7. Spatial Sampling, Aliasing, and Ambiguity Functions

In spatially sampled arrays, the chirp structure of the spherical-wave steering vector gives rise to a non-trivial, non-space-invariant bandlimit, which determines the minimum element spacing to avoid aliasing of the ambiguity function. For a ULA, the bandlimit is

$\mathbf{E}_{\text{inc}}(\mathbf{r})$0

Generalized Nyquist conditions follow, and if violated, ghost peaks appear in the array’s ambiguity function due to spatial bandfolding. The precise quantification of spatial sampling constraints in the Spherical Wavefront Regime (SWR) is central to array design in near- and intermediate-field scenarios (Monnoyer et al., 8 May 2025).

---

Key Papers Referenced:

• "Optimizing Sparse Antenna Arrays for Localization and Sensing using Vector Spherical Wave Functions" (Lafer et al., 19 Sep 2025)
• "Chirp-Based Aliasing Analysis of Arrays in the Spherical Wavefront Regime" (Monnoyer et al., 8 May 2025)
• "Expansion of Arbitrary Electromagnetic Fields in Terms of Vector Spherical Wave Functions" (Moreira et al., 2010)