Virtual Bistatic Sensing (VIBS) Overview

• Virtual Bistatic Sensing (VIBS) is a method that synthesizes bistatic radar links from distributed transmitters and receivers to enable high-resolution target localization and tracking.
• It leverages advanced nonlinear estimation and Bayesian tracking techniques, integrating independent angle and range measurements for sub-meter accuracy.
• VIBS is crucial for next-generation ISAC applications, underpinning innovations in ultra-massive MIMO, metasurface arrays, and flexible communication protocols for emerging 6G networks.

Virtual Bistatic Sensing (VIBS) is a paradigm within Integrated Sensing and Communication (ISAC) systems in which pairs of spatially distributed transmitters and receivers—or, more generally, antenna or sector pairs—are organized to emulate a bistatic radar geometry for high-accuracy target localization, tracking, and environment mapping. Rather than relying solely on dedicated bistatic ISAC hardware, VIBS synthesizes “virtual bistatic” links via the fusion of independently measured angles and ranges from communication infrastructure or programmable apertures. This enables sub-meter level performance for positioning and velocity estimation, exploiting both network diversity and signal processing innovations. VIBS frameworks are emerging as central features in ultra-massive MIMO, metasurface antenna arrays, asynchronous comms/radar systems, and extended 3GPP channel modeling standards.

1. Principle of Virtual Bistatic Sensing

VIBS operations rely on the formal fusion of measurements from virtual transmitter (TX) and receiver (RX) pairs, where any sector or antenna element can serve in either role per time or frequency slot. Each virtual bistatic link consists of a TX at known position $p_{\rm T}^i$ and an RX at $p_{\rm R}^i$, with the target at $p=[x,y]^T$ (2D) or $p=[x,y,z]^T$ (3D). For each link, relevant measurements include:

• TX bearing $$ϕ_{\rm T}^i = \arctan2(y-y_T^i, x-x_T^i) + \text{noise}$$
• RX bearing $$ϕ_{\rm R}^i = \arctan2(y-y_R^i, x-x_R^i) + \text{noise}$$
• Bistatic range $d_B^i = \|p-p_T^i\| + \|p-p_R^i\| + \text{noise}$

The fusion of these measurements forms the basis for nonlinear position and velocity estimation via maximum-likelihood methods (Bauhofer et al., 2024). The concept generalizes to programmable metasurfaces (Gavras et al., 12 Nov 2025), ultra-massive MIMO (Wan et al., 29 Dec 2025), and synthetic aperture (moving/steered) arrangements (Luo et al., 2024).

2. Measurement Models and Estimation Algorithms

VIBS systems implement nonlinear measurement models for state estimation: $$z_i = h_i(p) + η_i,\quad η_i \sim \mathcal{N}(0, R_i)$$ where $h_i(p)$ stacks the geometric transformations for bearings and range. For $N$ virtual bistatic links, the joint likelihood cost is

$$J(p) = \sum_{i=1}^N [z_i - h_i(p)]^T R_i^{-1} [z_i - h_i(p)]$$

The optimal position estimate $\hat{p}$ is obtained by minimizing $J(p)$, typically using Gauss–Newton or Levenberg–Marquardt techniques. The approach generalizes to 3D (by extending $p$) and incorporates velocity via fusion of Doppler measurements, as in antenna-pairwise FMCW echo models: $$\tau_{i,k} = \frac{1}{c} \|p - p_{t,i}\| + \frac{1}{c} \|p - p_r\|,\quad \nu_{i} = \frac{1}{\lambda}(\hat{u}_{i,t} + \hat{u}_{i,r})^T \vec{v}$$ Each DSA-subcarrier pair in UM-MIMO schemes acts as a virtual bistatic path, and the position/velocity estimation is solved by nonlinear least squares or closed-form linearizations (Wan et al., 29 Dec 2025).

3. Covariance Computation and Bayesian Tracking

For robust tracking, VIBS provides both position estimate $\hat{p}$ and its error covariance $P$, forming the input to a Bayesian tracker (e.g., Kalman Filter). Two covariance models are supported:

• Fixed (pre-tuned) covariance: Empirically derived via offline clustering/simulation, yielding $$P_{\text{fixed}} = \operatorname{diag}(\sigma_x^2, \sigma_y^2)$$; consistently yields strong performance absent measurement outliers.
• Dynamic (Hessian-based) covariance: Approximated as the inverse Hessian of $J(p)$ at $\hat{p}$,

$P_\text{dyn} \approx [H_J^T C^{-1} H_J]^{-1}$

where $H_J$ is the Jacobian of the nonlinear model. Alternate “converted measurement” covariances are obtainable via first-order Taylor expansion of geometric formulas (see Eq. 15 of (Bauhofer et al., 2024)).

The KF update equations incorporate $R_k = P_\text{fixed}$ or $P_\text{dyn}$; posterior covariance is computed via standard linear KF equations. Position/velocity RMSEs down to 0.25 m and 0.83 m/s are demonstrated in mmWave ISAC campus evaluations (Bauhofer et al., 2024).

4. Physical Aperture Models and Sensing Optimization

In metasurface-based VIBS architectures, the aperture is modeled as a coupled-dipole array with mutual coupling represented by

$$\mathbf{m} = I[\operatorname{diag}(\boldsymbol{\alpha})^{-1} - \mathbf{G}]^{-1} \mathbf{h}_f$$

where $\boldsymbol{\alpha}$ are per-element polarizabilities. Passivity constraints ensure each element’s response is physical: $$|\Im\{\alpha_n^{-1}\}| \geq \frac{k^3}{3\pi} + \frac{k^2}{8h}$$ Efficient design utilizes a Neumann-series approximation for rapid tuning. Performance bounds are derived via the Fisher information matrix and Cramér–Rao bounds (CRB) over multi-target coordinates (Gavras et al., 12 Nov 2025), and optimization solves for element tunings to minimize aggregate position error.

5. Signal Processing for Asynchronous Devices and Multipath Environments

VIBS can be realized even in clock-asynchronous, multipath-dominated environments using phase-invariance and differencing. By exploiting the fact that phase offsets from clock asynchrony and carrier drift are invariant across multipath components, algorithms subtract the LoS phase from each path and fit the resulting phase progressions to nonlinear least squares for Doppler and bearing estimation: $\Delta_i[k] = \mod_{2\pi}(\tilde{\phi}_i[k] - \tilde{\phi}_i[k-1]) \approx 2\pi T \cdot f_i^{\text{eff}} + \text{noise}$ This enables VIBS with commercial radios and multipoint reflections, requiring at least two static reflectors with distinct AoAs (Ventura et al., 2024). Median Doppler errors under 2% are reported at SNR ≥ 5 dB.

6. Channel Modeling and ISAC-VIBS Experimentation

The VIBS channel model extends 3GPP standards (TR 38.901) for bistatic configurations. It achieves compatibility with communication frameworks by:

• Doubling the number of channel clusters for weak-target capture, lowering the removal threshold from –25 dB to –50 dB.
• Supporting deterministic (e.g., ray-tracing) and statistical (randomized with spatial coherence) models for targets.
• Explicit spatial coherence via phase-correlation matrices $\mathbf{C}_n(m,m') = \exp[-(d_{m,m'}/D_c)^2]$.
• Adjusted path-loss exponents ($2n$), and shadow-fading variances ($\sqrt{2}\sigma$).

Experimental setups synthesize virtual array positions via electronic steering or physical movement, forming large synthetic-aperture data cubes for range–angle–Doppler processing. ISAC-VIBS frameworks preserve BER-vs-SNR curves under QPSK/16QAM, confirming full communications compatibility (Luo et al., 2024).

Key Parameter	Typical Value	Reference
Carrier Frequency	28 GHz	(Bauhofer et al., 2024)
Bandwidth	200–500 MHz	(Luo et al., 2024)
Range RMSE	0.05–0.25 m	(Bauhofer et al., 2024, Wan et al., 29 Dec 2025)
Velocity RMSE	0.02–0.83 m/s	(Bauhofer et al., 2024, Wan et al., 29 Dec 2025)
Number of VIBS links	3–16 (UM-MIMO, metasurface)	(Wan et al., 29 Dec 2025, Gavras et al., 12 Nov 2025)

7. Robustness and Practical Implementation

VIBS robustness in the face of spatial non-stationarity and hostile multipath is achieved by:

• Orthogonal signaling via chirps, antennas, or subcarriers (DSA/DSS) to isolate direct-path signals (Wan et al., 29 Dec 2025).
• Wideband and multistatic diversity to average out incoherent multipath and support cluster “birth–death.”
• Hardware-light metasurface architectures that time/frequency multiplex patterns to emulate bistatic TX/RX without duplicating RF chains (Gavras et al., 12 Nov 2025).

Implementation in ISAC networks entails calibration of TX/RX geometries, selection of update intervals compatible with maximum Doppler, and tuning system parameters (PRF, bandwidth, aperture element count) to balance resolution and complexity.

8. Future Directions and Application Domains

VIBS is central to future 6G applications requiring unified communications, localization, and sensing. Proposed extensions include dense-cellular deployments leveraging LM virtual links for large-scale tracking (Bauhofer et al., 2024), programmable metasurfaces with actively optimized element responses for environmental mapping (Gavras et al., 12 Nov 2025), and integrated asynchronous methods for COTS device-based radar (Ventura et al., 2024). This suggests further cross-standard developments, exploration of joint beamforming/spectrum utilization, and dual-use ISAC physical layers. The VIBS paradigm is foundational for next-generation high-resolution wireless sensing.