CSI-Free Low-Complexity Remote State Estimation over Wireless MIMO Fading Channels using Semantic Analog Aggregation
Minjie Tang ∗ , Photios A. Stavrou ∗ , and Marios Kountouris ∗ † {{}^{*}}^{\dagger} · 2025
Read CSI-Free Low-Complexity Remote State Estimation over Wireless MIMO Fading Channels using Semantic Analog Aggregation by Minjie Tang ∗ , Photios A. Stavrou ∗ , and Marios Kountouris ∗ † {{}^{*}}^{\dagger} (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-2501-14358.txt.
Introduction
Minjie Tang, Photios A Stavrou, and Marios Kountouris estimate a system's state without knowing the channel. Their 2025 paper targets remote monitoring over wireless MIMO links, where classic Kalman filtering demands perfect channel knowledge and heavy computation.
Their scheme skips channel estimation entirely. Sensors send the surprising part of their readings, the gap between measurement and expectation, merged over the air in analog form. A constant-gain filter, tuned offline by a staged approximation method, does the estimating at low cost. The team proves a stability condition in closed form using drift analysis. Simulations show the lean scheme beating several standard approaches, which the authors offer for sensor networks that cannot afford channel tracking.
Catalog record
| CREATOR | Minjie Tang ∗ , Photios A. Stavrou ∗ , and Marios Kountouris ∗ † {{}^{*}}^{\dagger} |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-14358
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
CSI-Free Low-Complexity Remote State Estimation over Wireless MIMO Fading Channels using Semantic Analog Aggregation Minjie Tang ∗ , Photios A. Stavrou ∗ , and Marios Kountouris ∗ † {{}^{*}}^{\dagger} Abstract In this work, we investigate low-complexity remote system state estimation over wireless multiple-input-multiple-output (MIMO) channels without requiring prior knowledge of channel state information (CSI). We start by reviewing the conventional Kalman filtering-based state estimation algorithm, which typically relies on perfect CSI and incurs considerable computational complexity. To overcome the need for CSI, we introduce a novel semantic aggregation method, in which sensors transmit semantic measurement discrepancies to the remote state estimator through analog aggregation. To further reduce computational complexity, we introduce a constant-gain-based filtering algorithm that can be optimized offline using the constrained stochastic successive convex approximation (CSSCA) method. We derive a closed-form sufficient condition for the estimation stability of our proposed scheme via Lyapunov drift analysis. Numerical results showcase significant performance gains using the proposed scheme compared to several widely used methods. I Introduction Remote state estimation has gained widespread interest recently [ 1 ] , with extensive applications in areas such as robotic control and autonomous navigation. A typical remote state estimation system consists of a physical dynamic plant , multiple geographically scattered sensors , and a remote state estimator , as shown in Fig. 1 . The wireless sensors measure the instantaneous internal states of the dynamic plant and deliver the state measurements to the remote state estimator over a wireless network. The remote state estimator generates the timely state estimation for the dynamic plant based on the received noisy measurements from the sensors. The wireless network between the remote state estimator and the sensors induces various impairments, such as fading and noise, which significantly degrade the estimation performance of the dynamic system. Fig. 1: Typical architecture of a remote state estimation system over a wireless network. Remote state estimation over a wireless network presents significant challenges. First, many existing works on remote state estimation overly simplify the wireless communication channels within the systems. For example, [ 2 ] utilizes an additive white Gaussian noise (AWGN) channel model, while [ 3 , 4 ] explores independent and identically distributed (i.i.d.) on-off failures in the wireless networks. However, these models overlook practical channel impairments, and directly applying such solutions to general fading channels may lead to estimation instability. Second, most existing works on remote state estimation assume dedicated radio resource allocation for multiple sensors [ 5 ] , resulting in poor spectral efficiency as sensor numbers grow. To address this, shared radio resource allocation for geographically dispersed sensors is preferred. Coordination of transmissions can be achieved using multiple access schemes such as ALOHA [ 6 ] and non-orthogonal multiple access (NOMA) [ 7 ] . However, these schemes face challenges with large sensor networks due to collision bottlenecks, and error propagation in the successive interference cancellation (SIC) process at the remote state estimator. Analog aggregation protocols [ 8 ] , which enable simultaneous transmission of sensor measurements without coordination, have recently been proposed to enhance spectral efficiency. However, these protocols rely on accurate CSI for state estimation, requiring pilot signal transmissions from each sensor for channel estimation at the remote state estimator [ 9 ] . This significantly increases communication overhead and introduces channel estimation noise [ 10 ] , which can undermine system stability. Recent advancements in semantic communications, as highlighted in [ 11 , 12 , 13 ] , emphasize the importance of transmitting relevant and useful application-specific information to reduce communication overhead. In the context of remote state estimation, this involves sending tailored sensor data rather than raw measurements. For instance, [ 14 ] investigates local Kalman filtering with central consensus, where local estimates are transmitted instead of raw data, potentially degrading performance due to measurement correlations. This underscores the need for further exploration of semantic communication strategies in remote state estimation. In this work, we introduce a novel CSI-free, low-complexity remote state estimation algorithm over wireless MIMO fading channels using semantic analog aggregation. The contributions of this work are summarized as follows: • Semantic Aggregation over Wireless MIMO Fading Channels. Analog aggregation [ 8 ] improves spectral efficiency and reduces latency in mission-critical applications but incurs substantial overhead due to the transmission of raw sensor data and pilot signals. To address this, we propose a semantic aggregation scheme in which sensors transmit only the discrepancy between predicted and actual measurements, leveraging predictions fed back from the remote state estimator. This approach preserves information integrity while drastically reducing communication overhead. • CSI-Free Low-Complexity State Estimation Solution. Existing remote state estimation methods that rely on Kalman filtering [ 3 , 4 , 5 , 8 ] require complex computations and perfect CSI. In contrast, we propose a constant-gain filtering algorithm that simplifies computation and eliminates the need for CSI. This approach avoids the transmission of pilot signals from the sensors and bypasses channel estimation at the remote state estimator, streamlining the entire process. • Estimation Stability Analysis. Using Lyapunov stability theory [ 15 ] , we derive a closed-form sufficient condition for ensuring estimation stability and propose a CSSCA-based offline algorithm to efficiently optimize the constant filtering gain. Notation: Uppercase and lowercase boldface letters denote matrices and vectors, respectively. The operators (\cdot)^{T} and \Tr(\cdot) represent the transpose and trace of a matrix, respectively. \mathbf{0}_{m\times n} and \mathbf{0}_{m} denote an m\times n matrix and an m\times m matrix with all elements equal to 0 , respectively. \mathbf{I}_{S} represents the S\times S identity matrix. \mathbb{R}^{m\times n} , \mathbb{S}_{+}^{m} , \mathbb{S}^{m} , \mathbb{Z}_{+} , and \mathbb{R}_{+} denote the set of m\times n real matrices, the set of m\times m positive definite matrices, the set of m\times m positive semi-definite matrices, the set of positive integers, and the set of positive real numbers, respectively. ||\mathbf{A}|| , \mathbf{A}\|_{F} and ||\mathbf{a}|| represent the spectral norm of a matrix \mathbf{A} , the Frobenius norm of a matrix \mathbf{A} and the Euclidean norm of a vector \mathbf{a} , respectively. [\mathbf{A}]_{i,j} denotes the i -th row and j -th column of the matrix \mathbf{A} . [\mathbf{a}]_{i} denotes the i -th element of the vector \mathbf{a} . II System Model In this section, we introduce the system model for the dynamic plant, wireless sensors, wireless communication channels, and remote state estimator. II-A Dynamic Plant Model We consider a discrete-time system with S\in\mathbb{Z}_{+} state variables. The physical plant state \mathbf{x}(t)\in\mathbb{R}^{S\times 1} evolves according to first-order coupled linear difference equations: \mathbf{x}(t+1)=\mathbf{A}\mathbf{x}(t)+\mathbf{w}(t),\,\,\,\,t=0,1,2,\ldots, (1) where \mathbf{A}\in\mathbb{R}^{S\times S} is the plant dynamics and \mathbf{w}(t)\sim\mathcal{N}(\mathbf{0}_{S\times 1},\mathbf{W}) is the plant noise with zero mean and finite noise covariance matrix \mathbf{W}\in\mathbb{S}^{S}_{+} . The initial plant state value \mathbf{x}(0)=\mathbf{x}_{0}\in\mathbb{R}^{S\times 1} . II-B Wireless Sensor Model We consider a system consisting of M\in\mathbb{Z}_{+} wireless sensors jointly monitoring the real-time plant state \mathbf{x}(t) . Each sensor has N_{t}\in\mathbb{Z}_{+} transmission antennas and shares radio resources for high spectral efficiency. The connection topology of the wireless sensors defines the connectivity between the plant and sensors. Definition 1 (Connection Topology of Sensors) The connection topology of (wireless) sensors is defined by a set of real matrices, \left\{\mathbf{C}_{1},\mathbf{C}_{2},\ldots,\mathbf{C}_{M}\right\}\in\mathbb{R}^{N_{t}\times S} , where each \mathbf{C}_{m} represents the connection between the dynamic plant and the m -th sensor. Specifically, an element [\mathbf{C}_{m}]_{i,s} is non-zero