Joint Sensing-Communications Design

• Joint sensing-communications design is the integrated co-optimization of radar sensing and wireless communications using shared spectral, hardware, and spatial resources.
• It employs full-duplex operation, MIMO, and wideband waveform techniques to simultaneously achieve high data throughput and precise target resolution.
• Key trade-offs in rate, resolution, and interference management are addressed through advanced optimization methods like alternating convex programs and deep unfolding algorithms.

Joint sensing-communications design refers to the rigorous co-optimization and physical integration of wireless communications and radar sensing within a unified transceiver platform, employing shared spectral, hardware, and spatial resources. This paradigm, also labeled as JCAS or ISAC, underpins advanced sixth-generation (6G) cellular architectures and related paradigms, encompassing full-duplex operation, multi-antenna systems, multi-user multiplexing, wideband waveform engineering, resource allocation, optimization under hardware constraints, and robust interference management. The joint design aims to simultaneously meet high-throughput communication requirements and high-resolution sensing metrics, explicitly quantifying and balancing trade-offs in rate, reliability, resolution, and latency.

1. Architecture and Signal Model Fundamentals

Modern JCAS architectures exploit a shared array platform capable of transmitting and receiving both communication and radar signals on overlapping time-frequency domains. The most comprehensive full-duplex JCAS topology consists of a base station (BS) with collocated transmit and receive arrays (MIMO, $N_t, N_r$), $K_n$ downlink half-duplex (HD) users, $K_u$ uplink HD users (also serving as bistatic radar illuminators), and support for both monostatic and bistatic radar operations. All signals—downlink and uplink user data, monostatic radar echoes, and bistatic echoes—are co-propagated using common hardware and real-time resources (Sheemar et al., 2023).

The composite signal model explicitly accounts for:

• Self-interference (SI) at the receiver: $y_{SI}(t) = H_{SI} x_{BS}(t) + n_{SI}(t)$, with $H_{SI}$ capturing SI channel after isolation.
• Downlink user reception: $$y_{DL,k}(t) = h_{BD,k}^H x_{BS}(t) + \sum_{i}g_{UL \rightarrow DL,k,i}^H x_{UL,i}(t) + n_{DL,k}(t)$$.
• Uplink multi-user reception: $$y_{UL}(t) = \sum_{i}H_{UL,i}x_{UL,i}(t) + H_{SI} x_{BS}(t) + n_{BS}(t)$$.
• Monostatic radar returns: $$r_{mono}(t) = \sum_{\ell} \alpha_{\ell} a_r(\theta_{\ell}) a_t^H(\theta_{\ell}) x_{BS}(t-\tau_{\ell}) e^{j2\pi f_{D,\ell}t} + n_r(t)$$.
• Bistatic radar echoes: $$r_{bi}(t) = \sum_{\ell} \beta_{i,\ell} H_{BS \leftarrow i,\ell} x_{UL,i}(t-\tau_{i,\ell}) e^{j2\pi f_{D,i,\ell}t} + n_r(t)$$.

Indicator: All radiated signals traverse the same channel and multipath environment, with mutual interference, convolution, and scattering processes that must be accurately modeled in both the physics and statistical domains (Sheemar et al., 2023).

2. Waveform and Resource Design Principles

JCAS systems require waveforms that simultaneously maximize data throughput and yield fine delay/Doppler resolution for target estimation. Wideband OFDM or chirp-like sequences are fundamental, obeying $B \gg$ data-rate bandwidth to enable $\Delta R = c/(2B)$ range resolution for radar, while still supporting modulation for communication (Sheemar et al., 2023).

Key design metrics:

• Ambiguity function $\chi(\tau, f)$: must be sharply peaked at $(\tau, f) = (0, 0)$ to yield precise range/Doppler localization and exhibit off-diagonal flatness to maintain orthogonality for communication multiplexing.
• Matched filtering and 2D FFT: applied across the time/pulse grid for each angle bin $\theta$ to extract $(\tau, f_D)$ pairs, facilitating simultaneous high-rate data link and multi-aspect radar imaging.

Multi-carrier and subcarrier-selective JCAS (for MIMO-OFDM) further partition subcarriers into roles, exploiting a subset for joint functions and reserving clean carriers for communications to enlarge the trade-off region: e.g., using $J\ll K$ subcarriers for sensing can boost total rate by $>60\%$ at constant beampattern fidelity (Nguyen et al., 2023).

3. Self-Interference Cancellation and Hardware Constraints

Full-duplex JCAS demands rigorous SI mitigation across passive, analog, and digital domains:

• Passive isolation: cross-polarization and physical separation can suppress SI by $30-50$ dB.
• Analog cancellation: RF-domain signals are subtracted with tunable canceller circuits (additional $40-60$ dB).
• Digital cancellation: baseband estimation and subtraction addresses residual SI after ADC.
• Residual SI model: $$P_{res} = E[\|(H_{SI} - \hat{H}_{SI}) x_{BS}\|^2] \approx Tr[(E[(H_{SI} - \hat{H}_{SI})(H_{SI} - \hat{H}_{SI})^H]) R_x]$$; the combined efficacy is critical to operating close to the thermal noise floor (Sheemar et al., 2023, Liu et al., 2022).

Hardware challenges include phase noise, amplifier nonlinearity, wideband circulator complexity at mmWave/THz, and sampling/switching power. Energy-efficient multi-beam analog arrays (MBAA) using fixed lens or Butler matrices provide discrete beamspace codebooks with minimal per-beam power overhead, enabling compositional beam synthesis and rapid angle scanning without expensive digital hardware (Wu et al., 2022).

4. Joint Optimization: Power, Beamforming, and Utility Functions

Joint sensing-communications optimization is formalized as constrained utility maximization:

• Objectives: maximize total sum-rate ($$R_{sum} = \sum_k \log_2(1 + \text{SINR}_{DL,k}) + \sum_i \log_2(1 + \text{SINR}_{UL,i})$$) and radar mutual information ($I_{radar} = \log\det(I + R_{radar} H_r H_r^H/N_0)$).
• Constraints: total radiated power $Tr(W W^H) \leq P_{max}$, SI $\leq \epsilon_{SI}$, user SINR bounds $\gamma_{DL,k}, \gamma_{UL,i}$, radar SNR requirements $\eta$ (Sheemar et al., 2023).

Weighted sum trade-off and multi-objective fairness (e.g., $\alpha$-fair utilities) are used to balance communication rates against sensing accuracy (Cramer-Rao Bound or mutual information). Typical optimization employs alternating convex programs, Riemannian manifold methods, or deep unfolding for fast and interpretable algorithmic convergence (Krishnananthalingam et al., 2023, Nguyen et al., 2024).

In dynamic environments (e.g., for mobile or time-varying channels), Lyapunov drift-plus-penalty methods control per-slot resource allocations, maintaining long-term reliability and dynamically adjusting beamformers to meet average constraints on SINR and radar SNR (Zakeri et al., 18 Mar 2025).

5. Trade-Offs and Performance Analysis

Joint design explicitly quantifies trade-offs among rate, resolution, and resource allocation:

• Rate vs. resolution: communication rate $R_{DC}=B_{comm} \log_2(1 + \text{SNR}_{comm})$ increases with $B_{comm}$; radar range resolution $\Delta R = c / (2B_{radar})$ improves with $B_{radar}$; total $B$ is shared.
• Numerical example: integration of multi-bistatic echoes can reduce angle estimation MSE by an order of magnitude at moderate user counts, illustrating multi-perspective synergy (Sheemar et al., 2023).
• Subcarrier partitioning: selective use of a subset for sensing preserves the beampattern while boosting communications sum-rate (Nguyen et al., 2023).
• Reliability/Spectral efficiency: joint design in sparse THz scenarios yields strictly positive correlation between reliability and spectral efficiency; no fundamental trade-off appears when both are bandwidth-limited (Chaccour et al., 2021).

Computation-efficient deep-unfolding approaches match optimal trade-off Pareto fronts at dramatically lower complexity and runtime ($30\times$ speedup over branch-and-bound for constant-modulus waveform design) (Krishnananthalingam et al., 2023).

6. Multi-User, Multi-Cell, and Environment-Aware Extensions

JCAS scales to multi-cell MISO and networked settings. Joint block-level precoder designs minimize sensing CRB and maximize min-SINR across users and cells. Semidefinite relaxation (SDR) and alternating optimization are used to realize beamformers; neglecting inter-cell reflections degrades performance by up to $5$ dB CRB loss at high SINR, while CoMP coordination achieves $2\times$ lower RCRB and $2$ dB SINR gain (Babu et al., 2024).

Environment-aware IRS deployment with dynamic beamforming leverages offline channel knowledge maps (CKM) for coverage optimization. SCA-based relax-and-bound algorithms minimize deployment costs for IRSs and BS operational costs, supporting per-point quality-of-service constraints for both sensing and communication sites (Chen et al., 5 Sep 2025).

7. Practical Considerations and Future Directions

Key implementation factors include:

• CSI acquisition: hierarchical channel estimation and blind source separation for overlapping comm/sensing echoes (Sheemar et al., 2023).
• Security: SI leakage risks require artificial noise subspace masking methods (Sheemar et al., 2023).
• Waveform flexibility: index modulation, dual-domain OFDM-delay-Doppler designs, and pilot-multiplexed schemes facilitate orthogonal multiplexing and efficient resource utilization (Elbir et al., 2024, Hua et al., 2022).
• Hardware integration: green JCAS via MBAA, low-rate ADC arrays via multi-subband quasi-perfect sequences, and co-design for THz (Mao et al., 2021, Wu et al., 2022).

Research continues in robust deep-unfolding optimization, distributed analog canceller networks for FD, dynamic multi-cell JCAS architectures, environment-adaptive IRS deployments, and scalable multi-user extensions. The field maintains an active discourse on trade-off optimality, hardware scaling, and real-time deployment challenges, with abundant scope for innovation as 6G and XR connectivity advance.