A Poincaré Lower Bound Approach for Performance Trade-offs in MIMO ISAC Systems with Blockage
Mohammadreza Bakhshizadeh Mohajer ∗ , Luca Barletta ∗ , Daniela Tuninetti † , Alessandro Tomasoni ∗∗ , Daniele Lo Iacono ∗∗ , and Fabio Osnato ∗∗ · 2025
Read A Poincaré Lower Bound Approach for Performance Trade-offs in MIMO ISAC Systems with Blockage by Mohammadreza Bakhshizadeh Mohajer ∗ , Luca Barletta ∗ , Daniela Tuninetti † , Alessandro Tomasoni ∗∗ , Daniele Lo Iacono ∗∗ , and Fabio Osnato ∗∗ (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-03707.txt.
Introduction
Mohajer and co-authors study in 2025 the trade off between sensing and data links in one shared wireless system. Such joint systems must split radio resources between watching the world and sending data. The sensing error is hard to bound when the channel can be blocked or take odd values.
They offer a new lower bound for sensing error that works in those hard cases. With it, they map an inner range of rates and sensing grades that can be reached on a fading channel with noise and block risk. They also give a plan that nears the best sensing grade as conditions grow large, which helps design of joint systems.
Catalog record
| CREATOR | Mohammadreza Bakhshizadeh Mohajer ∗ , Luca Barletta ∗ , Daniela Tuninetti † , Alessandro Tomasoni ∗∗ , Daniele Lo Iacono ∗∗ , and Fabio Osnato ∗∗ |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-03707
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A Poincaré Lower Bound Approach for Performance Trade-offs in MIMO ISAC Systems with Blockage Mohammadreza Bakhshizadeh Mohajer ∗ , Luca Barletta ∗ , Daniela Tuninetti † , Alessandro Tomasoni ∗∗ , Daniele Lo Iacono ∗∗ , and Fabio Osnato ∗∗ Abstract Characterizing the performance trade-offs between sensing and communication subsystems is essential for enabling integrated sensing and communication systems. Various metrics exist for each subsystem; however, this study focuses on the ergodic capacity of the communication subsystem. Due to the complexity of deriving the sensing mean square error (MSE) and the inapplicability of the Bayesian Cramér-Rao Bound to channels with discrete or mixed distributions, this work proposes a Poincaré lower bound on the sensing MSE to address these issues. An achievable inner bound for the rate-sensing trade-off in a fading multiple-input multiple-output channel with additive white Gaussian noise and blockage probability is established. In addition, a strategy that is asymptotically optimal for sensing is provided. ∗ Politecnico di Milano, 20133 Milano, Italy. Email: \{ mohammadreza.bakhshizadeh, luca.barletta \} @polimi.it † University of Illinois Chicago, 60607 Chicago, IL, USA. Email: [email protected] ∗∗ STMicroelectronics Srl, 20864 Agrate B.za, Italy. Email: firstname . lastname @st.com I Introduction The 6G mobile communication system has emerged as a significant focus of global research, encompassing a range of innovative application scenarios, including smart cities, intelligent transportation, smart manufacturing, and smart homes [ 1 ] . Precise sensing is expected to evolve from an auxiliary function to a fundamental service among the notable features of 6G, thereby adding an additional dimension of capability to the network [ 2 , 3 , 4 ] . There are various performance metrics for each individual system and based on their perspective. On one hand, the communication system is designed to maximize the information rate using an optimal transmitted signal while having knowledge of the channel. On the other hand, the sensing subsystem solves the problem of estimating the parameters of the channel while it has knowledge of the transmitted signal [ 5 ] . Future 6G vehicle-to-everything (V2X) is foreseen to explore millimeter wave (mmWave) [ 6 ] . Furthermore, the anticipated utilization of millimeter-wave (mmWave) and terahertz (THz) frequencies for sensing in 6G and next-generation wireless networks is foreseen [ 4 ] . These frequencies are subject to increased losses and are more susceptible to blockage [ 7 ] . I-A Previous Work Motivated by integrated sensing and communication (ISAC) systems and propagation challenges in high frequencies, we study the performance trade-off between the two subsystems. The performance of the communication subsystem is typically evaluated by the ergodic capacity. Due to the challenging evaluation of the mean square error (MSE), which is the typical performance metric of the sensing subsystem, the Bayesian Cramér-Rao bound (BCRB) on the MSE has been adopted by previous work [ 5 ] . In contrast to the classic Cramér-Rao bound (CRB), which depends on the actual value of the parameter being estimated, the BCRB utilizes the a priori distribution of the parameter and provides a global bound [ 8 ] . However, the issue with BCRB is that it is generally not tight and does not apply to scenarios involving discrete or mixed distributions [ 9 ] . To address this issue, in our previous work [ 10 ] , we proposed an alternative lower bound on the MSE of channel estimation based on a Poincaré-type inequality derived in [ 9 ] . Specifically, in [ 10 ] we computed a lower bound on the sensing MSE for multiple-input multiple-output (MIMO) additive Gaussian channels with Rayleigh fading and blockage events. I-B Contributions In this work, we compute an achievable sensing-rate region along with an outer bound for a MIMO ISAC system. Furthermore, we suggest an achievable strategy that is asymptotically sensing optimal for high sensing signal-to-noise ratio (SNR). Finally, through numerical simulations, we evaluate the inner bound and discuss the results. I-C Paper Organization In Sec. II , we introduce the system model and the performance metrics for both sensing and communication subsystems and we define the structure of sensing-rate inner bound. The achievable strategies are studied in Sec. III . The numerical evaluation of the achievable boundary and a discussion on the results are provided in Sec. IV . Finally, Sec. V concludes the paper. I-D Notation Throughout the paper, deterministic scalar quantities are denoted by lowercase letters, random vectors are denoted by uppercase sans serif letters, deterministic vectors are denoted by bold lowercase letters, and random matrices by bold uppercase letters (e.g. x,\mathsf{X},\mathbf{x},\mathbf{X} ). \langle\cdot,\cdot\rangle denotes the inner product. For a matrix \mathbf{A} , \mathbf{A}^{\mathsf{T}} , \mathbf{A}^{\dagger} , \mathbf{A}^{-1} , |\mathbf{A}| , and \text{tr}(\mathbf{A}) denote the transpose, the Hermitian transpose, the inverse, the determinant and the trace of the matrix \mathbf{A} , respectively. (\mathbf{A})^{+}=(\mathbf{A}^{\mathsf{T}}\mathbf{A})^{-1}\mathbf{A}^{\mathsf{T}} denotes the pseudoinverse of matrix \mathbf{A} . For matrix \mathbf{A}\in\mathbb{R}^{n\times m} , \text{vec}(\mathbf{A})=[\mathbf{a}_{1}^{\mathsf{T}},\mathbf{a}_{2}^{\mathsf{T}},\dots,\mathbf{a}_{m}^{\mathsf{T}}]^{\mathsf{T}}\in\mathbb{R}^{nm\times 1} , where \mathbf{a}_{i} is the i -th column of \mathbf{A} , is the vectorization operator. \mathbf{I}_{k} is the identity matrix of dimension k , \bf 0 is the column vector of all zeros, \Gamma(\cdot) indicates the gamma function, and \delta(\cdot) denotes the Kronecker delta. The smallest eigenvalue and the smallest singular value are denoted by \lambda_{\textrm{min}}(\mathbf{A}) and \sigma_{\min}(\mathbf{A}) , respectively. \succeq denotes the semidefinite ordering and the Kronecker product is denoted by \otimes . \nabla_{\mathbf{x}}(\cdot) , \mathbf{J}_{\mathbf{x}}(\cdot) , and \mathbb{H}_{\mathbf{x}}(\cdot) denote the gradient, the Jacobian matrix, and the Hessian matrix with respect to \mathbf{x} . The minimum MSE (MMSE) of estimating the random vector \mathsf{X} given observation \mathsf{Y} is defined as \text{mmse}(\mathsf{X}|\mathsf{Y})\triangleq\mathbb{E}\left[\|\mathsf{X}-\mathbb{E}[\mathsf{X}|\mathsf{Y}]\|^{2}\right], (1) where \mathbb{E}[\cdot] denotes the expectation operator and \|\cdot\| is the Euclidean norm. The {\mathsf{Var}}\left[\cdot\right] is defined as \displaystyle{\mathsf{Var}}\left[\mathsf{X}\right] \displaystyle\triangleq\mathbb{E}\left[\|\mathsf{X}-\mathbb{E}[\mathsf{X}]\|^{2}\right]. (2) We also consider the following complex to real mapping for column vectors and matrices \displaystyle\overline{\mathbf{x}}=\begin{bmatrix}\text{Re}({\mathbf{x}})\\ \text{Im}({\mathbf{x}})\end{bmatrix},\,\overline{\mathbf{X}}=\begin{bmatrix}\text{Re}({\mathbf{X}})&-\text{Im}({\mathbf{X}})\\ \text{Im}({\mathbf{X}})&\text{Re}({\mathbf{X}})\end{bmatrix}. (3) II System Model and Achievable Inner Bound II-A System Model Let us consider the general MIMO ISAC system model \displaystyle\mathbf{Y}_{i} \displaystyle=\mathbf{H}_{i}\mathbf{X}+\mathbf{Z}_{i},\quad i\in\{\rm c,\rm s\} (4) where \mathbf{Y}_{\rm c}\in\mathbb{C}^{N_{\rm c}\times T} and \mathbf{Y}_{\rm s}\in\mathbb{C}^{N_{\rm s}\times T} are the communication and sensing received signals, \mathbf{X}\in\mathbb{C}^{M\times T} is the transmitted dual-function signal for both communication and sensing tasks, \mathbf{H}_{\rm c}\in\mathbb{C}^{N_{\rm c}\times M} denotes the communication channel matrix, and \mathbf{H}_{\rm s}\in\mathbb{C}^{N_{\rm s}\times M} is the target response matrix. The channel matrices have the following fading distribution \displaystyle f_{\mathbf{H}_{i}}(\mathbf{x})=(1-\alpha_{i})\delta(\mathbf{x})+\alpha_{i}\mathcal{CN}\left(\mathbf{x};{\bf 0},\sigma_{\mathsf{H}_{i}}^{2}\mathbf{I}_{N_{i}M}\right),i\in\{\mathrm{c},\mathrm{s}\} (5) where 1-\alpha_{i} , for \alpha_{i}\in(0,1] , is the probability that receiver i experiences a blockage event that lasts T channel uses; both channels vary every T symbol in an independent and identically distributed (i.i.d.) manner; and the noise vectors \mathbf{Z}_{i} are assumed to have i.i.d. circularly symmetric, zero mean, complex Gaussian components with variance \sigma_{i}^{2} , for i\in\{\mathrm{c},\mathrm{s}\} . The sample covariance matrix \mathbf{R}_{x} of the transmitted signal will play a central role in both the analysis of the sensing system’s performance and in imposing an average-power constraint P_{0} for the communication problem, i.e., \displaystyle{\text{tr}}\left({\mathbb{E}}\left[\mathbf{R}_{x}\right]\right)\leq MP_{0},\quad\mathbf{R}_{x}\triangleq T^{-1}\mathbf{X}\mathbf{X}^{\dagger}\in\mathbb{C}^{M\times M}. (6) We assume \mathbf{X} to be known at the sensing receiver, which observes \mathbf{Y}_{\rm s} and aims to estimate the target response matrix \mathbf{H}_{\rm s} . The communication subsystem aims to transmit as much information as possible through optimal \mathbf{X} while the transmitter and the communication receiver have full knowledge of \mathbf{H}_{\rm c} . The performance of the communication subsystem