Introduction
In the evolution toward 6G wireless networks, Joint Sensing and Communication (JSC) has emerged as a cornerstone technology. Traditionally, JSC research focuses on how communication signals—such as those from a base station—can be dual-purposed for radar-like sensing of the environment. However, a provocative alternative exists: "Communication via Sensing." This paradigm flips the script, exploring how a receiver's sensing capabilities (e.g., radar, LiDAR, or cameras) can be used to decode information intentionally embedded in the physical state changes of a target device.
The Problem: Communicating Through Movement
Consider a low-power IoT device or a micro-robot that lacks the hardware to transmit traditional radio frequency (RF) signals. If a nearby sensing infrastructure can track this device's position, orientation, or even temperature, the device can communicate by deliberately transitioning between these physical states.
Unlike traditional digital communication, where "bits" are mapped to electromagnetic phases or amplitudes, this system maps information to physical attributes. This introduces unique constraints:
- Physical Continuity: A device cannot teleport; it must move through a sequence of intermediate states.
- Energy Constraints: Every physical transition incurs a cost (e.g., battery consumption for movement).
- Sensing Noise: The receiver's sensors are imperfect, introducing errors into the "read" process.
Characterizing the fundamental capacity—the maximum achievable data rate—of such a channel requires a departure from standard information theory, as the "input" is governed by physical transition rules and energy budgets.
Methodology: Modeling the Finite-State Channel
The authors model this system as an irreducible Finite-State Channel (FSC). To overcome the computational complexity of calculating capacity over infinite sequences, the research employs several key mathematical strategies:
State Pruning and Sequence Decomposition
To simplify the underlying Markov chain, the framework prunes the state-space to include only the most cost-effective actions between any two states. A long sequence of movements is then mathematically decomposed into a weighted sum of "cycles" (loops returning to a starting state) and a short direct path.
Single-Letter Formulation
Calculating the $N$-letter capacity (the limit as the sequence length $N$ approaches infinity) is typically intractable. By focusing on the properties of the decomposed cycles, the authors convert the problem into a "single-letter" optimization. This allows for the calculation of capacity based on the statistical frequency of cycles rather than the specific path taken.
Duality and Minimax Optimization
Using the dual form of the capacity problem and applying Sion’s minimax theorem, the researchers established a tight upper bound on the information rate. This bound accounts for the noise in the sensing process, modeled as a Binary Symmetric Channel (BSC).
Key Insights
The research yields several critical insights into the nature of physical-state communication:
- Modulation via State: Physical coordinates or orientations serve as "constellation points." The efficiency of this modulation is directly tied to the device's ability to transition between these points.
- The Cost-Information Trade-off: There is a strict relationship between the average cost budget ($\Gamma$) and capacity. If the energy budget is below a specific threshold ($\Gamma_{min}$), communication is impossible. Conversely, once the budget exceeds a certain point ($\Gamma_{max}$), the capacity saturates as the system hits its unconstrained physical limit.
- Cycle Dominance: In long-term communication, the specific trajectory matters less than the cycles the device traverses. This insight is the mathematical "engine" that makes the capacity calculation feasible.
Results and Performance
The proposed framework was validated through numerical simulations of a robot moving within a grid. The results demonstrate that the derived single-letter upper bound is remarkably tight, nearly overlapping with the achievable lower bound across various noise levels (from crossover probabilities of $p=0$ to $p=0.2$). For simplified two-state systems, the authors successfully derived closed-form expressions for the capacity limit, providing a clear benchmark for future system designs.
Limitations and Future Directions
While providing a robust foundation, the study identifies several areas for expansion:
- State Initialization: The current model assumes the initial state is known to both parties, though its influence diminishes over long sequences.
- Complexity in Transitions: The pruning method assumes a single "best" path between states, which may not apply to high-dimensional environments with multiple variables.
- Scalability: Future work will focus on scaling these models to accommodate hundreds of possible states (e.g., high-resolution drone coordinates) and exploring multi-user scenarios where a single sensor reads multiple devices simultaneously.
Conclusion
"Communication via Sensing" offers a viable path for ultra-low-power devices to transmit data by leveraging the existing sensing infrastructure of 6G networks. By establishing the information-theoretic limits of this approach, this research provides the mathematical tools necessary to design efficient, state-based communication protocols.
Summary updated