Entropy-Based Sensing Schemes for Energy Efficiency in Massive MTC

Sergi Liesegang,Antonio Pascual-Iserte,Olga Muñoz

DOI: https://doi.org/10.1109/LWC.2020.2984223

2024-12-07

Abstract:Machine-type communications (MTC) are crucial in the evolution of mobile communication systems. Within this context, we distinguish the so-called massive MTC (mMTC), where a large number of devices coexist in the same geographical area. In the case of sensors, a high correlation in the collected information is expected. In this letter, we evaluate the impact of correlation on the entropy of a set of quantized Gaussian sources. This model allows us to express the sensed data with the data correlation matrix. Given the nature of mMTC, these matrices may be well approximated as rank deficient. Accordingly, we exploit this singularity to design a technique for switching off several sensors that maximizes the entropy under power-related constraints. The discrete optimization problem is transformed into a convex formulation that can be solved numerically.

Signal Processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to improve energy efficiency by selectively turning off certain sensors in the massive machine - type communication (massive MTC) scenario while ensuring that the amount of information collected is as large as possible. Specifically, the paper focuses on how to utilize the high correlation of sensor data to reduce redundant information, thereby reducing energy consumption and prolonging the battery life of sensors in the same geographical area where a large number of sensors co - exist.

### Problem Background
In massive machine - type communication (mMTC), a large number of sensor devices will be deployed in the same area, and the data collected by these sensors are often highly correlated. If all sensors keep working and transmitting data continuously, it will lead to a great deal of energy waste. Therefore, how to effectively select which sensors should work and which can be turned off to maximize the amount of information while minimizing energy consumption has become an important research topic.

### Core Problem of the Paper
The core problem of the paper is to design an entropy - based sensing scheme. By selectively turning off some sensors, under the premise of meeting the power constraint, the entropy (i.e., the amount of information) of the collected information is maximized. Specifically:
- **Objective**: Under a given power constraint, select a set of sensors to make the entropy of the data they collect the largest.
- **Method**: Use the entropy of the quantized Gaussian source and the data correlation matrix to represent the sensor data, and solve the optimal sensor selection scheme through convex optimization methods.

### Mathematical Model and Optimization Problem
The paper uses the following mathematical models to describe the problem:
1. **Multivariate Entropy**:
\[
H(X^\Delta)=-\sum_{x^\Delta} p_{X^\Delta}(x^\Delta)\log_2 p_{X^\Delta}(x^\Delta)
\]
2. **Differential Entropy**:
\[
h(X)=-\int f_X(x)\log_2 f_X(x)\,dx
\]
3. **Approximation of Quantized Entropy**:
\[
eH(X^\Delta)\triangleq h(X)-M\log_2\Delta\leq H(X^\Delta)
\]
4. **Selection Matrix \(B\)** and the Corresponding Optimization Problem:
\[
b^* = \arg\max_{b\in\{0,1\}^M} eH(X_p^\Delta)\quad\text{s.t.}\quad\|f(b)\|\leq\delta
\]

### Solution
To simplify this discrete optimization problem, the paper proposes the following steps:
1. **Relax the Constraint Set**: Relax \(b\in\{0,1\}^M\) to \(b\in[0,1]^M\).
2. **Approximate the Objective Function**: Use the nuclear norm to approximate the rank function, and use the trace to replace the \(l_0\) quasi - norm.
3. **Solve the Convex Optimization Problem**: Eventually transform it into a convex optimization problem that can be solved by standard numerical methods.

### Conclusion
Through simulation results, the paper proves that the proposed scheme can retain most of the useful information while significantly reducing energy consumption. Especially in the massive MTC scenario, due to the high correlation of sensor data, the effect of this method is more obvious.