Coverage and Deployment Analysis of Narrowband Internet of Things in the Wild


The content in this webpage complements the paper “Coverage and Deployment Analysis of Narrowband Internet of Things in the Wild”, for which a preprint version can be found here: https://arxiv.org/abs/2005.02341. The paper is currently under revision in IEEE Communications Magazine (Internet of Things and Sensor Networks Series). The complete dataset will be made available soon.


Tool Description

Measurement System

For the NB-IoT measurements in Oslo and Rome, we used the Rohde&Schwarz (R&S) TSMA6 toolkit, together with an Exelonix Narrowband (NB) USB device and a global position-ing system (GPS) antenna.

Dataset Description

A description of the dataset attributes is available below:

Coverage

Available for LTE and NB-IoT:

  • Date, Time, UTC: Timestamp
  • Latitude, Longitude, Altitude: GPS coordinates
  • Speed: Moving speed
  • Heading: Movement direction (if available)
  • X.Sat: Number of available GPS satellites
  • TAC: Tracking Area Code
  • EARFCN, Frequency: Carrier frequency identifiers (anonymized)
  • MCC, MNC: Operators identifiers at country level (anonymized)
  • eNodeB.ID, (N)PCI, CI, cellID: Numerical identifiers for E-UTRAN Node Bs (eNBs) and cells (anonymized if unique)

Available for LTE:

  • BW: Bandwidth [MHz]
  • SymPerSlot: Number of symbols per slot (depends on the use of normal vs. extended OFDM cyclic prefix)
  • Power: Power received on the entire BW [dBm]
  • SINR: Signal-to-Interference-plus-Noise Ratio [dB]
  • RSRP: Reference Signal Received Power [dBm]
  • RSRQ: Reference Signal Received Quality [dB]
  • X4G.Drift: Time offset between cell and scanner [ns]
  • Sigma.4G.Drift: Standard Deviation on X4G.Drift
  • TimeOfArrival: Time of Arrival (ToA) from cell to scanner [s]
  • TimeOfArrivalFN: ToA Frame Number (FN). It allows to report TimeOfArrival on an absolute time scale, ToA_abs = TimeOfArrival + TimeOfArrivalFN * 10 [ms]

Available for NB-IoT:

  • NSINR.Tx0, NSINR.Tx1: Narrowband SINR for antenna ports Tx0 and Tx1 [dB]
  • NRSRP.Tx0, NRSRP.Tx1: Narrowband RSRP for antenna ports Tx0 and Tx1 [dBm]
  • NRSRQ.Tx0, NRSRQ.Tx1: Narrowband RSRQ for antenna ports Tx0 and Tx1 [dB]
  • RSSI.Tx0, RSSI.Tx1: Received Signal Strength Indicator (RSSI) for antenna ports Tx0 and Tx1 [dBm]
  • NSSS.Power: Narrowband Secondary Synchronization Signal (NSSS) Power [dBm]
  • NSSS.RSSI: NSSS RSSI [dBm]
  • NSSS.CINR: NSSS Carrier-to-Interference-plus-Noise Ratio [dB]

eNBs and cells

  • Name: eNB name (combines eNodeB.ID and PhyCellID)
  • Latitude, Longitude: Cell GPS coordinates 
  • PosErrorDirection: Cell directional error
  • PosErrorLambda1, PosErrorLambda2: Cell longitudinal and latitudinal positioning errors 
  • IsDirected: Binary indication, omnidirectional (0) vs. directional (1) cell
  • Direction: Direction identifier
  • Power: Cell Power read by the scanner when the cell position estimation is finalised [dBm]
  • MaxPowerUsedForTowerEstimationbyPE: maximum value of the cell received power used to estimate the cell location [dBm] 
  • TowerID: Tower identifier
  • MCC, MNC: Operators identifiers at country level (anonymized)
  • TAC: Tracking Area Code
  • PhyCellID, CellIdentity: Numerical identifiers for cells (anonymized if unique)
  • EARFCN: Carrier frequency (anonymized)

Dataset

A sample of the dataset consisting of one complete sub-campaign collected in Outdoor Driving (OD) scenario is available below:

sample data

Visualization


Coverage Maps
eNB Positioning

Paper Figures/Complementary Figures/Statistical Analyses


Paper Figure 1: Scatterplot between average distance to the eNB with highest RSRP and average distance to the nearest eNB for Op_1,N (left) and Op_2,N (right). Different colors represent different scenarios.

Paper Figure 2: Sub-campaign average coverage in terms of RSRP [dBm], grouped by scenario and divided by a combination of operator (Op_1,N, Op_2,N) and technology (LTE, NB-IoT).

To enforce and validate the above result, we perform additional statistical significance tests to pinpoint differences between the means of the distributions. We examine the following use cases.

  1. Operator Comparison: For a given scenario and technology, we show if there is significant difference between the two operators.
  2. Technology Comparison: For a given scenario and operator, we show if there is significant difference between the two technologies.
  3. Scenario ComparisonFor a given operator and technology, we show if there is significant difference between any combination of the four scenarios.
  • We perform the Kruskal-Wallis non-parametric analysis of variance test to assess whether data comes from the same distribution. For the scenario comparison, since we have more than two groups, we also leverage the Dunn’s test that is used to identify significant difference between the means of two or more distributions for a given confidence interval.
  • Null Hypothesis: There is no significant difference between the two distributions
  • Alternative Hypothesis: There is significant difference between the two distributions.
  • We report the p-values below — Significant codes: ‘*’ 0.05 , ‘**’ 0.01, ‘***’ 0.001
TechnologyDIIOWOD
LTE0.29060.0274 *0.24800.3581
NB-IoT0.15940.00033 ***0.11520.1439
Operator comparison
OperatorDIIOWOD
10.0021 **2.3e-09 ***0.0117 *1e-05 ***
20.0133 *1e-06 ***0.00078 ***1e-05 ***
Technology comparison
Technology/OperatorDI-IDI-OWDI-ODI-OWI-ODOW-OD
LTE/14.8e-08 ***6.4e-06 ***0.00041 ***0.360330.300550.17261
LTE/25.7e-07 ***6.4e-06 ***3.6e-06 ***0.36860.57520.6063
NB-IoT/15.7e-07 ***3.1e-05 ***3.1e-06 ***0.52990.52990.9434
NB-IoT/23.1e-06 ***1.1e-06 ***1.6e-07 ***0.0492 *0.0442 *0.8244
Scenario comparison

Complementary Figure 1: Network topology scenario for the in-band vs guard-band comparative analysis.

Paper Figure 3: Network deployment performance evaluation in terms of RSRP [dBm] for the in-band Op_1,I deployments.

Complementary Figure 2: Network deployment performance evaluation in terms of RSRP [dBm] for the guard-band Op_2,I deployments.

Likewise, we follow up with the statistical analysis by leveraging the Kruskal-Wallis test. We examine the following scenario

  • Technology ComparisonFor a given operator and R, we show if there is significant difference between any combination of the two technologies.

We report the p-values below — Significant codes: ‘*’ 0.05 , ‘**’ 0.01, ‘***’ 0.001

Operator100200300400500
13.8e-10 ***<2e-16 ***<2e-16 ***8.6e-07 ***0.0683
20.75540.06680.0169 *0.74740.9627
Technology comparison
%d bloggers like this: