Introduction
As wireless technology transitions toward the 6G era, the fundamental objectives of network design are expanding. Beyond the traditional goal of high-speed data transmission, future networks are envisioned to incorporate Integrated Sensing and Communication (ISAC). In this paradigm, radio signals serve a dual purpose: conveying digital messages and acting as radar-like probes to sense the physical environment. This paper establishes a rigorous theoretical framework for ISAC within multi-user environments, specifically focusing on the Multiple Access Channel (MAC).
The Challenge of Multi-User ISAC
In a standard communication system, the priority is maximizing the bit rate while maintaining low error probability. In ISAC, however, there is an inherent trade-off between communication throughput and sensing precision (distortion). Optimizing a signal for data often degrades its utility for environmental sensing, and vice versa.
While previous literature has explored this trade-off for single-user or broadcast scenarios, this work addresses the more complex Multiple Access Channel (MAC). In this setup, multiple transmitters (e.g., various sensors or mobile devices) send information to a single receiver (e.g., a base station) simultaneously. The complexity is heightened by the fact that these transmitters may have only partial or noisy observations of the environment and can receive feedback from the channel, allowing for intricate coordination and interaction.
Methodology: The SD-DMMAC Model
The authors model the system as a State-Dependent Discrete Memoryless Multiple Access Channel (SD-DMMAC). To map the achievable boundaries of this system, they utilize several advanced information-theoretic techniques:
- Markov Coding and Two-Stage Decoding: The researchers implemented a coding scheme that spans multiple time blocks. The decoder employs a "backward-forward" strategy, utilizing information from future blocks to refine the decoding of current messages.
- Generalized Feedback: The model assumes encoders can monitor channel feedback. This allows transmitters to align their environmental knowledge and coordinate their signals more effectively.
- Auxiliary Variable Splitting: A significant innovation in this work is the use of two distinct sets of auxiliary variables. The first set ($T$) is optimized for communication performance, while the second set ($V$) is dedicated exclusively to helping the decoder estimate the environmental state.
- Shannon Strategy: This approach allows transmit signals to adapt dynamically based on the causal side information—real-time environmental data—available at the encoders.
Key Theoretical Insights
Feedback as a Collaborative Tool
In a multi-user sensing context, feedback serves a purpose beyond simple error correction. It acts as a synchronization mechanism, allowing disparate encoders to "align" their observations. This collaborative approach results in a joint estimation of the environment that is superior to what any individual encoder could achieve in isolation.
The Role of "Sensing-Only" Variables
The study demonstrates that achieving optimal sensing performance requires specific information (the $V$-variables) that is decoded only after the primary message. While these variables do not contribute to the data rate, they are essential for minimizing "distortion"—the mathematical measure of sensing error.
Unified Architectural Modeling
The paper validates its broad mathematical framework by showing it can be simplified to match known results in specific architectures, such as:
- Monostatic-uplink ISAC: Where a base station functions as both a radar transceiver and a data receiver.
- Multi-modal Sensor Networks: Where distributed sensors with varying capabilities (e.g., motion and optical) cooperate to localize targets while transmitting data.
Results and Implications
The primary contribution of this research is Theorem 1, which defines the "achievable region" for the system through a set of 17 mathematical inequalities. These inequalities delineate the possible combinations of data rates ($R_0, R_1, R_2$) and sensing distortion ($D$).
For the specific case of monostatic-uplink ISAC, the authors prove that their proposed region is not merely an "inner bound" (a set of achievable rates) but actually coincides with the optimal Capacity-Distortion (C-D) region. Furthermore, the paper provides a quantitative analysis of rate splitting, showing how private messages should be divided into common and private components to maximize the benefits of the feedback loop.
Conclusion and Future Work
This work provides a foundational mathematical roadmap for 6G ISAC systems, yet several avenues for future research remain. Currently, the paper establishes an achievable inner bound; proving that these rates represent the absolute maximum possible (the outer bound or converse) for the general case remains an open challenge.
Additionally, the current model assumes all participants are honest. Future iterations of this framework must incorporate security and secrecy constraints to prevent sensing signals from being exploited by eavesdroppers. Finally, the authors suggest that 6G will require new definitions of "trustworthiness" in sensing, which may necessitate expanding the mathematical definition of distortion to include reliability and integrity metrics.
Summary updated