Highly Reconfigurable Pixel Antenna

• HRPA is an electronically controlled antenna featuring a sub-wavelength pixel grid that allows dynamic reconfiguration of radiating topology and feed positions.
• It employs RF switches (e.g., PIN diodes or MEMS) and multiport network theory to enable precise far-field pattern synthesis and robust multiuser performance.
• The architecture enhances sensing, MIMO communications, and wireless power transfer through advanced codebook optimization and efficient EM pattern control.

A Highly Reconfigurable Pixel Antenna (HRPA) is an electronically controlled antenna architecture in which the radiating surface is discretized into sub-wavelength elements ("pixels") interconnected via radio frequency (RF) switches. HRPA advances traditional pixel antenna designs by enabling simultaneous reconfiguration of both the radiating surface topology and feeding port positions, thus providing enhanced control over electromagnetic (EM) patterns, spatial degrees of freedom, and system-level performance in communication, sensing, and wireless power transfer domains. Recent research has demonstrated that HRPA architectures can drastically improve key metrics such as angle estimation accuracy, spectral efficiency, energy transfer, and multiuser throughput relative to conventional fixed or uniform arrays (Han et al., 19 Jan 2026).

1. Pixel Antenna Architecture and Fundamental Principles

HRPA structures consist of a planar radiating surface divided into a grid of $Q$ pixels (e.g., $5\times5$ or denser), fabricated on substrates such as Rogers 4003C. Each pixel can be shorted or opened to its adjacent neighbors through PIN-diode or MEMS switches, which are digitally programmed to achieve target connection patterns. In advanced HRPA, feeding flexibility is provided by configuring $M$ potential feed ports (typically located at pixel centers and routed through the ground plane), where $N$ ports are activated and the rest muted via digital switching (FPGA or MCU control) (Han et al., 19 Jan 2026). This joint reconfigurability allows not only the sculpting of aperture geometry (radiating topology), but also dynamic feed-position switching, which is essential for full-sphere angular sensing and uniform direction-finding performance.

The electromagnetic behavior is described by a multiport impedance matrix $Z$ (size $(M+Q)\times (M+Q)$) capturing all feed and loaded port interactions. Binary switch states $g_q\in\{0,1\}$ map to port loads: open-circuit ($g_q=1$) or short-circuit ($g_q=0$), with $Z_L^L(g)$ encoding the active switch configuration. Feeding-port impedance and radiated patterns are then computed as functions of $(F, g)$, enabling closed-form modeling of mutual coupling and pattern control.

2. Circuit Theory, Far-Field Pattern Synthesis, and EM Degrees of Freedom

The core modeling approach leverages multiport network theory. Active and passive ports are partitioned, and the effective input impedance for any feed configuration $(F,g)$ is given by:

$$Z_F(F, g) = Z_{AA} - Z_{AL}\cdot (Z_{LL} + Z_L^L(g))^{-1}\cdot Z_{LA}$$

The "loaded" open-circuit far-field pattern is synthesized as:

$$E_{oc,F}(F, g, \Omega) = E_A - \left[ E_M,\ E_L \right] \left[ \begin{array}{cc} Z_{MM} + Z_M^L & Z_{ML} \ Z_{LM} & Z_{LL} + Z_L^L \end{array} \right]^{-1} \begin{bmatrix} Z_{MA} \ Z_{LA} \end{bmatrix}$$

The mutually coupled polarization components $e_\Theta(F,g,\Omega),\ e_\Phi(F,g,\Omega)$ follow after accounting for source impedances and per-port efficiency:

$$E(F, g, \Omega) = E_{oc,F}(F, g, \Omega) \left(Z_0 + Z_F(F,g)\right)^{-1} \Lambda^{1/2}(F, g)$$

EM degrees-of-freedom (DoF) are determined by the rank of the open-circuit radiation matrix $E_{oc}$ (from SVD analysis), typically scaling with aperture size and pixel density (Han et al., 5 Dec 2025). HRPAs exploit these DoF to achieve low-correlated beam patterns, enabling enhanced spatial multiplexing and angle discrimination.

3. Codebook-Based Angular Sensing: CRLB-Driven Optimization

For sensing applications, HRPA enables codebook-based covering of the spatial domain. The 3D sphere is partitioned into $K$ disjoint angular sectors $\{\Omega_k\}$; for each sector, precomputed codewords $X_k = (F_k, g_k)$ specify optimal feed and switch configurations. Upon coarse angle determination, the appropriate $(F_k, g_k)$ is loaded, ensuring sector-dependent pattern diversity. The criterion for codeword selection is the minimization of the sector-wise worst-case Cramér–Rao lower bound (CRLB) on direction estimation:

$$\min_{F\subset\{1,\ldots,M\},\ g\in\{0,1\}^Q} \max_{\Omega\in\Omega_k} \sqrt{\operatorname{Tr}[C(F,g,\Omega)]}$$

where the CRLB matrix $C(F,g,\Omega)$ is derived from the Fisher Information Matrix incorporating polarization derivatives and null-space projectors (Han et al., 19 Jan 2026, Sams et al., 3 Dec 2025). Optimization proceeds via alternating search: genetic algorithm updates switch states ($g$), while greedy feed selection minimizes the CRLB. Multi-stage sector subdivision (warm-started over parent codes) enables compact codebooks.

Simulations reveal HRPA achieves $>50\%$ reduction in direction-finding error (CRLB) across all sphere sectors compared to aperture-matched uniform planar arrays (UPA), especially mitigating the UPA's pronounced endfire degradation (Han et al., 19 Jan 2026). Increasing the number of active feeds (up to $N=8$) improves performance, though mutual coupling introduces diminishing returns.

4. Exploiting EM Pattern Coding for ISAC, Beamspace MIMO, and Wireless Power Transfer

The HRPA pattern coder principle underpins a new class of integrated sensing and communication (ISAC), fluid antenna systems, MIMO spatial multiplexing, and wireless power transfer (WPT) frameworks (Zhang et al., 3 Dec 2025, Han et al., 5 Dec 2025, Sams et al., 3 Dec 2025, Chen et al., 12 Jan 2026). By rapidly cycling antenna-coder vectors, HRPA realizes instantaneous pattern switching—equivalent to physical movement—yielding rich sampling of electromagnetic channels.

Beamspace system models represent MIMO links as concatenations of low-rank EM pattern coders. For each transmitter, the pattern coder coordinates $w(s) = \Sigma V^\mathrm{H} i(s)$ form a compact basis via SVD of the aperture's open-circuit radiation modes. Multiuser, multi-antenna links benefit from codebook-driven pattern selection, maximizing sum-rate under total power and hardware constraints (Li et al., 11 Dec 2025, Sams et al., 3 Dec 2025). Alternating optimization algorithms combine fractional programming for digital beamformers and SEBO or genetic search for antenna coders.

In WPT, binary or continuous antenna coding is synergistically paired with transmit beamforming and nonlinear rectifier modeling. Alternating convex approximation and codebook clustering (K-means, Lloyd) allow scalable solution of joint beamforming and antenna coding, yielding >15 dB average output DC power gains over fixed arrays (with continuous coding adding another 6 dB) (Chen et al., 12 Jan 2026, Chen et al., 16 Dec 2025).

5. Hardware Implementation and Experimental Performance

HRPA implementations use high-density pixel grids (e.g., $5\times5$ or greater) with PIN-diode or MEMS switches controlled by FPGAs/MCUs. Feed port switching (up to $M$ candidate locations) is achieved via digitally routed multiplexers (Han et al., 19 Jan 2026, Debogovic et al., 2014, Zhang et al., 2024). Bias networks (DC lines, RF chokes, capacitors) are carefully isolated from RF paths for robust switch actuation.

Experimental validation shows multi-state HRPAs maintain $S_{11}<-10$ dB matching across all configurations; far-field gain/efficiency (typically >6 dBi, >80%) remains stable. Pattern-covariance matrices conform closely to Bessel-function models, confirming rich spatial diversity. Switching times are μs-scale; measured pattern agreement between simulation and hardware typically falls within 1–2 dB, with minor deviations due to component tolerances and switch losses.

Large-scale HRPA architectures (over 1000 pixels) are feasible via monolithic MEMS or advanced PCB processes, with scalability limited primarily by bias routing and switch density (Debogovic et al., 2014). MEMS-based designs achieve near-zero power consumption, sub-10 μs switching, and >70% array efficiency (metasurfaces) in Ku-band applications.

6. Design Trade-Offs, Complexity, and Codebook Construction

The HRPA configuration space grows exponentially with pixel count ($2^Q$). Exhaustive search is infeasible; alternating codebook approaches are standard. Lloyd-type algorithms, genetic optimization, and hierarchical codebooks reduce selection complexity while maintaining near-optimal performance for multiuser and ISAC scenarios (Li et al., 11 Dec 2025, Chen et al., 12 Jan 2026). Balanced hardware designs select pixel density ($Q$), feed count ($M,N$), and switch technology to optimize for pattern DoF, insertion loss, and control complexity.

For practical deployment, codebook sizes of $M=8$–16 deliver most of the theoretical gains; hierarchical search achieves substantial CPU-time reduction at minor sum-rate loss. Pattern orthogonality and correlation error thresholds guide switch-count reduction strategies and coder selection.

7. Applications and System-Level Impact

HRPA architectures are adopted in advanced sensing (full-sphere angle estimation, radar, ISAC), fluid antenna systems, MIMO spatial multiplexing, multiuser beamforming, wireless power transfer, and electromagnetically aware transceivers for 6G and beyond. In angular sensing, HRPA outperforms fixed UPA arrays, achieving nearly uniform estimation performance across all directions (Han et al., 19 Jan 2026). In multiuser communication, HRPA-enabled arrays (REMAA/FC-REMAA) approach movable-antenna (MMA) performance, with power-loss gaps reduced below 3.25% given sub-$0.1\lambda$ pixel intervals (Chen et al., 1 Apr 2025). In energy transfer, gains of >15 dB over fixed arrays, and codebook-driven designs achieve 40% improvement at fractional complexity (Chen et al., 12 Jan 2026, Chen et al., 16 Dec 2025).

HRPA’s capacity to jointly reconfigure radiating geometry and feeding ports marks a fundamental advance over prior architectures. It delivers programmable EM diversity, robust codebook-based pattern selection, efficient beamspace channel modeling, and system-level power/spectral-efficiency gains with minimal additional RF hardware complexity. The HRPA paradigm is central to next-generation EM-aware and software-defined antenna systems.