Tri-Hybrid Beamforming Architecture

• Tri-hybrid beamforming is a three-layer structure integrating digital, analog, and electromagnetic precoding to enhance degrees of freedom and overall performance.
• It employs joint optimization techniques such as fractional programming, manifold optimization, and alternating optimization to balance communication and sensing objectives.
• Implementations use reconfigurable antennas, metasurfaces, and dynamic radiation control to achieve significant improvements in sum rate, SCNR, and energy efficiency.

Tri-hybrid beamforming architectures generalize hybrid beamforming by introducing a third layer—electromagnetic (EM) or radiation-domain beamforming—besides conventional baseband digital and RF analog stages. This three-pronged approach, often coupled with reconfigurable antennas, metasurface arrays, or pattern-selective radiation center placement, aims to maximize the degrees of freedom (DoF), spectral efficiency (SE), energy efficiency (EE), and robustness in advanced wireless systems such as integrated sensing and communication (ISAC), extra-large MIMO, and mmWave/THz networks. Contemporary realizations span digital/RF/EM design, codebook and manifold optimization, dynamic antenna placement, and tri-timescale control strategies.

1. Tri-Hybrid Beamforming System Model

The tri-hybrid beamforming system embodies three cascaded precoding stages:

1. Digital Baseband Precoder ($\mathbf{F}_{\rm BB}$): Performs multi-user MIMO precoding and interference mitigation. Typically, $$\mathbf{F}_{\rm BB}\in\mathbb{C}^{N_{\rm RF}\times K}$$, for $N_{\rm RF}$ RF chains and $K$ data streams.
2. Analog RF Precoder ($\mathbf{F}_{\rm RF}$): Realized via phase-shifters; shapes array gain and performs constant-modulus beamsteering, often constrained such that $|[\mathbf{F}_{\rm RF}]_{i,j}|=1$.
3. EM/Radiation-Domain Precoder ($\mathbf{F}_{\rm EM}$): Models the spatial radiation pattern configuration, implemented by reconfigurable antennas (ERAs), dynamic metasurfaces, pinching antennas, or selected radiation centers. Frequently, $\mathbf{F}_{\rm EM}$ is block-diagonal, with antenna-specific pattern vectors (e.g., spherical harmonics or RC selection).

The generalized transmit signal is

$$\mathbf{x} = \mathbf{F}_{\rm EM}\,\mathbf{F}_{\rm RF}\,\mathbf{F}_{\rm BB}\,\mathbf{s}$$

Radiation and EM design injects additional free parameters per radiating element, thereby significantly expanding the system's effective DoF beyond traditional hybrid architectures (Chen et al., 16 Oct 2025).

2. Key Degrees of Freedom and DoF Scaling

Tri-hybrid beamforming enhances DoF through the radiation-domain stage:

• Digital DoF: $2K N_{\rm RF}$, where $K$ is the number of data streams/users.
• Analog DoF: $N_T N_{\rm RF}$ (phase shifter network).
• EM/RC DoF: For ERAs, $2N_T(T-1)$ for spherical-harmonic patterns (with truncation $T$), or variable-level per-antenna weights for metasurfaces, pinching positions, or RC selection.

The total real-valued DoF is

$$D_{\text{tri-HBF}} = 2K N_{\rm RF} + N_T N_{\rm RF} + 2N_T(T-1)$$

In contrast, conventional hybrid is limited to $2K N_{\rm RF} + N_T N_{\rm RF}$, restricting how finely beams can be shaped, sidelobes can be suppressed, and nulls can be formed (Chen et al., 16 Oct 2025). This expansion is central to the architecture's performance advantage across communication and sensing metrics.

3. Optimization Frameworks and Solution Methods

Tri-hybrid beamforming designs typically solve a joint optimization problem balancing communication rate and sensing gain (e.g., SCNR):

$$\max_{\mathbf{F}_{\rm EM},\mathbf{F}_{\rm RF},\mathbf{F}_{\rm BB}} \tilde\beta\,R_c + \beta\,\eta$$

subject to power, constant-modulus, and physical/antenna constraints.

The solution methodology is modular:

• Fractional Programming (FP): Transforms tricky ratio/objective terms into tractable forms using auxiliary variables; e.g., for SINR and SCNR (Chen et al., 16 Oct 2025, Li et al., 21 Aug 2025).
• Manifold Optimization (MO): Used for unit-modulus constraints, e.g., analog phase-shifter weights lie on complex circle manifolds. Riemannian conjugate-gradient is common (Chen et al., 16 Oct 2025, Zhao et al., 18 Nov 2025).
• Alternating Optimization (AO): Decouples variables into sequential subproblems: fix radiation patterns, optimize digital/analog; fix BB/RF, optimize EM/radiation; iterate to convergence (Chen et al., 16 Oct 2025, Jr. et al., 28 May 2025).
• Metaheuristics: Evolutionary algorithms (e.g., SHADE) optimize discrete radiation center selection or pinching antenna positions (Zhao et al., 18 Nov 2025).
• Closed-form Updates: In designs with RC or EM stage selections, quadratic transforms and constrained matrix inversions enable rapid per-iteration updates (Chen et al., 16 Oct 2025, Li et al., 21 Aug 2025).

Performance bottlenecks typically reside in matrix inversion ($O(N_T^3)$) and per-antenna EM or RC updates ($O(K\,T)$), though specialized methods (DQTFP, LDTFP) accelerate the inner loop with closed-form steps (Li et al., 21 Aug 2025).

4. Hardware Realizations and Architectures

The tri-hybrid concept subsumes several physical realizations:

Architecture	Radiation Design	Typical Physical Components
ERA-ISAC (Chen et al., 16 Oct 2025)	Spherical harmonics	Tunable load networks, PIN diodes
Metasurface/DMA (Fang et al., 22 Jan 2026)	Metasurface weights	Dynamic metasurface arrays, programmable pixels
Pinching Antenna PASS (Cheng et al., 2 Nov 2025, Zhao et al., 18 Nov 2025)	PA position optimization	Dielectric waveguides, mobile PA elements, MEMS actuators
RC Reconfigurable Array (Li et al., 21 Aug 2025)	Radiation center selection	Binary array selection, RC switching
Multi-timescale (Liu et al., 5 Mar 2025)	Dynamic pattern alignment	Reconfigurable array elements, tri-layered control systems

Radiation layer elements (ERAs, DMAs, PASS, RCs) must be controllable in real time, typically via tunable hardware (PIN diodes, MEMS), programmable varactors, or software-defined pattern-selective switching. Calibration and mutual coupling effects pose practical challenges, especially for spherical-harmonic or block-diagonal RC patterns. Tri-timescale frameworks decouple update rates to minimize pilot overhead and computational cost (Liu et al., 5 Mar 2025).

5. Performance Benchmarks and Trade-Offs

Tri-hybrid architectures yield substantial improvements over conventional hybrid and fully-digital counterparts:

• Sum Rate: Achieves up to 10 dB gain in weighted S&C performance at low power, and up to 8 bps/Hz sum rate at −20 dBm, compared to 5 bps/Hz for classic hybrid (Chen et al., 16 Oct 2025).
• Sensing Quality (SCNR/Power): Realizes 15 dB SCNR, approaching optimal radar illumination (Chen et al., 16 Oct 2025, Fang et al., 22 Jan 2026).
• Energy Efficiency: Gains of 2–3× over hybrid and digital baselines due to lower RF/PS chain count and passive element scaling (Fang et al., 22 Jan 2026, Zhao et al., 18 Nov 2025, Li et al., 21 Aug 2025).
• Beam Pattern Precision: Narrow main lobes and deep nulls via digital+analog+EM joint control (Chen et al., 16 Oct 2025, Cheng et al., 2 Nov 2025).
• Robustness to CSI Errors: Sensing/communication gain persists even under up to 20% channel estimation error (Zhao et al., 18 Nov 2025).
• Pilot Overhead Reduction: Tri-timescale frameworks reduce training by focusing slow updates on EM/radiation layer and fast updates on digital baseband (Liu et al., 5 Mar 2025).

A plausible implication is that tri-hybrid systems can sustain near-fully-digital spectral efficiency with ~50% fewer RF chains, scaling energy and spatial gain linearly with passive array size (Fang et al., 22 Jan 2026, Jr. et al., 28 May 2025).

6. Typical Applications and Domain Extensions

Tri-hybrid beamforming forms the backbone of several wireless paradigms:

• Integrated Sensing and Communication (ISAC): Jointly optimizes communication rate and SCNR (Chen et al., 16 Oct 2025, Fang et al., 22 Jan 2026).
• mmWave/THz Extra-Large MIMO: Harnesses low-power metasurface antennas for XXL arrays, addresses spatial non-stationarities (Fang et al., 22 Jan 2026, Liu et al., 5 Mar 2025).
• Pinching-Antenna Systems: Real-time adaptation to LoS path geometry, coverage regions (Cheng et al., 2 Nov 2025, Zhao et al., 18 Nov 2025).
• Energy-Constrained Deployments: EE maximization for massive IoT, vehicular, and access scenarios (Li et al., 21 Aug 2025, Xu et al., 2019).
• Robust Beamforming: Offers resilience to hardware errors, mutual coupling, and pattern uncertainty (Fang et al., 22 Jan 2026, Zhao et al., 18 Nov 2025).

Extensions include multi-target/multi-cell ISAC, wideband/multi-carrier design, low-overhead channel estimation for dynamic arrays, and robust beamforming under patterned or stochastic uncertainties (Chen et al., 16 Oct 2025, Liu et al., 5 Mar 2025).

7. Open Challenges and Research Directions

Current research highlights several unsolved problems:

• Realizable Pattern Constraints: Spherical harmonic truncations may yield nonphysical patterns; enforcing amplitude/phase bounds and mutual coupling limits is necessary (Chen et al., 16 Oct 2025).
• Hardware Calibration and Speed: Fast reconfiguration (<1 ms), calibration of dense ERA/RC arrays, and distributed control across layers remain active areas (Chen et al., 16 Oct 2025, Jr. et al., 28 May 2025).
• Integrated Multi-objective Design: Joint optimization of radiation pattern and digital beamforming codebook design (Chen et al., 16 Oct 2025, Li et al., 21 Aug 2025).
• Learning-Based Algorithms: Potential for neural or reinforcement learning approaches for ultra-fast layer updates and feedback control, especially in distributed environments (Han et al., 2021).
• Robustness and Latency: Dynamic adaptation in mobile/multi-user scenarios demands low-latency reconfiguration of the EM and analog layers.

This suggests continued emphasis on closed-form low-complexity algorithms (e.g., LDTFP over DQTFP), scalable RC/ERA hardware platforms, and systematic calibration under realistic deployment conditions (Li et al., 21 Aug 2025, Zhao et al., 18 Nov 2025).

---

Notable References:

• ERA-aided ISAC joint digital/analog/EM optimization (Chen et al., 16 Oct 2025)
• DMA-based tri-hybrid design with alternating close-form algorithm (Fang et al., 22 Jan 2026)
• Pinching antenna PASS-enabled tri-hybrid capacity scaling (Cheng et al., 2 Nov 2025, Zhao et al., 18 Nov 2025)
• RC selection in reconfigurable arrays for SE and EE (Li et al., 21 Aug 2025)
• Tri-timescale control for pilot overhead minimization (Liu et al., 5 Mar 2025)
• Tri-hybrid concept in large-array cmWave/6G MIMO transceivers (Jr. et al., 28 May 2025)
• SDR hardware validation and waveform-domain precoding (Xu et al., 2019)