Source-linked AI summary
People effects on IoT indoor wireless channel characterization
Millena Michely de Medeiros Campos, Mateus de Oliveira e Mattos, Rafael da Silva Macedo, Alvaro Augusto Machado de Medeiros, Wellerson Viana de Oliveira, Vicente Angelo de Sousa Junior
TL;DR
The paper addresses limited evidence on how people affect indoor wireless channels below 1 GHz, an issue for IoT propagation characterization. It analyzes 900-MHz narrowband measurements and finds that occupancy changes fading-distribution parameters, with Nakagami fitting the measured small-scale variations best.
Problem
Indoor channel propagation varies with environmental features and people, while evidence on people’s effects below 1 GHz in high-occupancy IoT settings is limited.
Source: §1 · p. 2Method
The study characterizes 900-MHz narrowband measurements by removing large-scale variation, normalizing the envelope, and fitting fading distributions with MLE and KS error analysis.
Source: §1 · p. 2, §3 · pp. 3–4Results
People significantly influenced small-scale fading distribution parameters, Nakagami fit the scenario best, and estimated Doppler shift was unaffected near the envelope mean.
Source: §5 · pp. 7–8Takeaways & Limitations
For this propagation scenario, Nakagami was the most suitable distribution for describing small-scale variations.
Source: §5 · p. 7
Abstract
from arXiv · showhide
Wireless communication under 1 GHz is suitable for Internet of Things (IoT) applications due to larger coverage capability with less power consumption. Bearing in mind that people and elements contained in the environment can cause variations in the channel, this paper aims to evaluate the effect of the presence of people on a 900-MHz indoor narrowband wireless channel, as we characterize the small-scale phenomena. With the increase in the number of people, a greater variation in the communication channel was noticed, which is reflected in the parameters of the probability distributions used in the characterization of the random part of the signal. In addition, second-order statistics were used to analyze the data and an adherence test was applied to confirm the behavior of the signal in relation to the distributions.
1 Introduction
Indoor wireless channels vary with environmental features and people, so free-space or outdoor models may not describe indoor propagation. This paper addresses limited evidence on people’s effects below 1 GHz by studying 900-MHz narrowband measurements in a busy IoT-like setting.
1 Introduction
- Indoor propagation varies with walls, furniture, and people, making outdoor or free-space models unsuitable in some indoor environments.This variability can occur even over shorter transmitter–receiver distances.Source: §1 · p. 2
- Prior studies examined indoor propagation below 1 GHz and people’s effects at 2 GHz, but evidence for high-occupancy IoT scenarios below 1 GHz was lacking.Wi-Fi gesture-recognition studies also showed that recognizing people’s influence on channel signals requires a large command-learning database.Source: §1 · p. 2
- The study evaluates people’s effects using 900-MHz narrowband measurements in the UFJF refectory as a high-density IoT operating scenario.The paper situates this work amid growing IoT demand and narrowband technologies using the 900-MHz ISM band.Source: §1 · p. 2
2 Measuring System
The measuring system combines a 900-MHz RF-generator transmitter with a computer-controlled SDR receiver. The receiver’s signal reception was linear across the tested transmitting-power range.
2 Measuring System
- A 900-MHz continuous-wave signal at 0 dBm was transmitted by an Agilent RF generator, while a USRP N210 SDR and computer captured measurements.The receiver used a WBX 50–2200 MHz daughterboard at a 10 MHz sample rate.Source: §2 · p. 2
- Matching vertically polarized omnidirectional antennas were mounted 1.60 m above ground to capture multipath effects in a wearable- or sensor-like configuration.The computer controlled measurement capture and storage.Source: §2 · p. 2
- Reception was linear from -80 to 5 dBm of transmitting-power variation in the USRP N210 calibration graph.The receiver was validated using an RF generator connected by cable with a 10 dB attenuator at 900 MHz.Source: §2 · pp. 2–3, Figure 2 · p. 3
3 Data measurement and processing methods
The study compares empty and crowded refectory measurements, extracts normalized small-scale fading, and fits candidate fading distributions to the observed envelope data.
3 Data measurement and processing methods
- Measurements compared an empty refectory with a crowded one, with transmitter and receiver 35 m apart in line of sight.The refectory was operating at half capacity, with an approximate daily flow of 1,200 people; counts were grouped into average-occupancy categories.Source: §3 · p. 3
- A sliding-window filter removed large-scale variations, then the measured envelope was divided by its RMS value to obtain the normalized small-scale signal.The resulting envelope was used to analyze small-scale fading variations.Source: §3 · p. 3
- The empirical CDF of small-scale fading samples was analyzed using Rice, Nakagami, and Weibull distributions; the cited passages state Rayleigh equivalence conditions for Rice and Nakagami, but not Weibull.Rice models environments with LOS propagation and is a particular case of Rayleigh when k = 0. Nakagami models multipath clusters and is equivalent to Rayleigh when m = 1; Weibull models multipath fading with non-linearities.Source: §3 · p. 4
- Maximum likelihood estimation determined each distribution’s parameters, and the KS error measured the distance between empirical and target CDFs.The KS error was calculated as the supremum of the difference between the empirical CDF and target CDF.Source: §3 · p. 4
4 Results and Discussions
Increasing occupancy is associated with higher channel variability and changing small-scale fading distribution parameters. The Nakagami distribution fits the measured data best, while second-order results depend on the signal-level threshold.
4 Results and Discussions
- Average received power differs by almost 10 dB between the empty and crowded scenarios despite line-of-sight reception.Figure 4 shows received-power samples over 450 seconds for the two scenarios.Source: §4 · p. 4, Figure 4 · p. 5
4.1 Analyzing the complete dataset
- Pooling all Scenario 2 measurements makes theoretical CDFs difficult to fit, motivating analysis by occupancy.The paper attributes the difficulty mainly to diverse measurement conditions.Source: §4 · p. 5, Figure 5 · p. 5
4.2 Analyzing the influence of people
- Distribution parameters decrease as restaurant occupancy rises, indicating greater channel variability with more people.Table 1 reports this inverse dependence across the tested campaigns.Source: Table 1 · p. 5, §4 · p. 5
- The Nakagami distribution fits the empirical data best; Campaign 1 KS errors are smaller at higher occupancies, with the same tendency in other campaigns.The paper notes KS error sensitivity to heavy-tailed distributions when measurements do not span the full distribution domain.Source: §4 · pp. 5–6, Figure 6 · p. 6
- For 50-person occupancy intervals, Nakagami remains the best fit; Figure 7 illustrates empirical and theoretical CDFs for 250–300 people.The interval-specific distribution parameters are reported in Table 2.Source: §4 · p. 6, Figure 7 · p. 6, Table 2
4.3 Second-order analysis
- For measurement data grouped into classes of 50 people, as occupancy increases, fd becomes more stable even at lower thresholds, because envelope variation intensifies and yields more level crossings there.Source: §4 · p. 7
5 Conclusions
The study finds that occupancy significantly influences fitted small-scale fading parameters, while estimated Doppler shift changes little near the envelope mean.
5 Conclusions
- Measurements used low-cost SDR equipment and included up to a few hundred people.Source: §5 · p. 7
- Across three days of measurements, the Nakagami distribution best described small-scale variations, and people significantly influenced fitted distribution parameters.Source: §5 · p. 7
- Despite considerable effects on Nakagami m, estimated Doppler shift varied little across occupancy near the envelope mean.Source: §5 · pp. 7–8
- Future work will evaluate antenna height and positioning, as well as combined indoor and outdoor scenarios.Source: §5 · p. 8