Alexander Shurakov,Margarita Ershova,Abdukodir Khakimov,Anatoliy Prikhodko,Evgeny Mokrov,Vyacheslav Begishev,Galina Chulkova,Yevgeni Koucheryavy,Gregory Gol'tsman
Abstract:Beam tracking is an essential functionality of millimeter wave (mmWave, 30-100 GHz) and sub-terahertz (sub-THz, 100-300 GHz) 5G/6G systems. It operates by performing antenna sweeping at both base station (BS) and user equipment (UE) sides using the Synchronization Signal Blocks (SSB). The optimal frequency of beam tracking events is not specified by 3GPP standards and heavily depends on the micromobility properties of the applications currently utilized by the user. In absence of explicit signalling for the type of application at the air interface, in this paper, we propose a way to remotely detect it at the BS side based on the received signal strength pattern. To this aim, we first perform a multi-stage measurement campaign at 156 GHz, belonging to the sub-THz band, to obtain the received signal strength traces of popular smartphone applications. Then, we proceed applying conventional statistical Mann-Whitney tests and various machine learning (ML) based classification techniques to discriminate applications remotely. Our results show that Mann-Whitney test can be used to differentiate between fast and slow application classes with a confidence of 0.95 inducing class detection delay on the order of 1 s after application initialization. With the same time budget, random forest classifiers can differentiate between applications with fast and slow micromobility with 80% accuracy using received signal strength metric only. The accuracy of detecting a specific application however is lower, reaching 60%. By utilizing the proposed technique one can estimate the optimal values of the beam tracking intervals without adding additional signalling to the air interface.
What problem does this paper attempt to address?
### What problem does this paper attempt to solve?
This paper aims to solve the problem of beam - tracking frequency optimization in the millimeter - wave (mmWave, 30 - 100 GHz) and sub - terahertz (sub - THz, 100 - 300 GHz) frequency bands in 5G/6G systems. Specifically, the author focuses on how to optimize the frequency of beam - tracking according to the type of application currently used by the user equipment (UE) to improve the resource utilization and performance of the system.
#### Background and problem description
In 5G/6G systems, beam - tracking is a key function to ensure beam alignment between the base station (BS) and user equipment. However, current standards do not clearly specify the optimal frequency of beam - tracking events, which depends on the micromobility properties of the application currently used by the user. Due to the lack of a clear application - type signal - transmission mechanism, the base station cannot directly obtain which application the user is using, making it difficult to make optimal beam - tracking decisions.
#### Solution
To solve this problem, the author proposes a method for remotely detecting applications based on the received signal - strength pattern. The specific steps are as follows:
1. **Measurement activity**: The author conducts a multi - stage measurement activity in the 156 - GHz sub - terahertz frequency band and collects the received - signal - strength data of four typical smartphone applications (voice call, video viewing, VR viewing, game).
2. **Statistical analysis**: Through statistical analysis of the time series of the received - signal - strength, the author finds that there are significant differences in the micromobility behaviors of different applications.
3. **Classification technique**: Using the classical Mann - Whitney test and machine - learning (ML) classification techniques (such as random forest), the author can distinguish between application categories with fast and slow micromobility properties.
#### Main contributions
- **Measured data**: Provides measured data of the received - signal - strength patterns of different types of applications in the 156 - GHz frequency band.
- **Classification method**: Proposes a remote - application - classification method based on the Mann - Whitney test and ML techniques, which can accurately identify application categories within a short time after beam - tracking.
- **Performance evaluation**: The results show that the Mann - Whitney test can distinguish between fast and slow application categories with 95% confidence within 1 second; the random - forest classifier has an accuracy of over 80% within 5 seconds.
Through this method, the base station can estimate the optimal beam - tracking interval without adding additional signaling, thereby improving the resource utilization and performance of the system.
### Formula display
To further explain the impact of micromobility behavior on beam - tracking, we can introduce some formulas. For example, the received - signal - strength \(P_r(t)\) can be expressed as:
\[P_r(t)=10\log\left(\frac{V_r}{S_u}\right)+30\]
where:
- \(V_r\) is the diode - response voltage,
- \(S_u = 100\,\text{V/W}\) is the diode - detector responsivity at 156 GHz.
In addition, the regression - slope coefficient \(\beta\) can be calculated by the least - squares method (LSF):
\[ \beta=\frac{\sum_{i = 1}^{n}(t_i-\bar{t})(P_r(t_i)-\bar{P_r})}{\sum_{i = 1}^{n}(t_i-\bar{t})^2}\]
These formulas help to understand the change of the received - signal - strength over time and provide a theoretical basis for the classification algorithm.