Outdoor
0 dB additional lossCalculation for outdoor sensor points.
How we calculate
the radio prediction.
How well can a sensor reach the planned gateways? The model estimates the link margin for Martinfeld from terrain, buildings and your settings.
Radio planning →868 MHzSensor → Gateway
746 buildingsOfficial LoD2 model
approx. 20 mSpacing between result grid points
From site to link margin
Determine heights
Terrain comes from the official DGM1 digital terrain model; roofs come from TLBG LoD2 building data. The gateway antenna is placed at the modelled roof or terrain elevation plus the selected mounting height. The sensor is placed at terrain elevation plus the selected sensor height.
Calculate the radio path
For each site, terrain and roofs are sampled along the connection. Simplified diffraction loss for obstacles is added to free-space path loss. Each gateway site is calculated separately.
Compare margins
Receiver sensitivity, planning margin and any additional indoor loss are subtracted from the estimated received level. With a remaining margin of at least 0 dB, a grid point is considered covered in the model.
More gateways: better margin or more coverage?
The best link margin counts at each grid point. The received powers of multiple gateways are not added together.
- Coverage
- Proportion of validly calculated grid points that reach at least one gateway with sufficient margin.
- Multiple coverage
- Proportion of valid grid points that reach at least two gateways. The overlap view and tapping a point show which gateways are involved.
- Additional benefit
- A gateway can cover new points, create a second reception option or improve the margin. Once everything is covered, the coverage percentage cannot increase further.
The percentages describe the outdoor grid in the village model – not population, building floor area or a guaranteed service coverage rate.
Outdoor, indoor and deep indoor
Indoor
Starting value: 15 dBA uniform additional loss applied to the same outdoor grid, adjustable during planning.
Deep Indoor
Starting value: 30 dBIndoor loss plus further attenuation for more heavily shielded sites, also adjustable.
15 and 30 dB are freely chosen planning assumptions, not measured or standardised values. Wall materials, floors and basements are not calculated individually. Building footprints remain excluded as reception locations.
What does the antenna comparison tell us?
The comparison tests 0, 2, 3, 5, 6, 8, 9 and 12 dBi using the same sites and grid. Each gain value is applied to all gateways. The recommendation is the smallest tested value that meets the selected coverage target. If none reaches the target, no suitable recommendation is given.
Gain is applied as a constant receive gain. Actual radiation patterns, vertical sidelobes and installation effects are not yet included. The recommendation is therefore a model comparison, not a definitive antenna or purchasing recommendation.
Reading the colours correctly
- 30 dB and above
- 15 to below 30 dB
- 0 to below 15 dB
- −10 to below 0 dB
- Below −10 dB
These colours apply to link margin. The overlap and gateway views use their own legends. Transparent areas represent excluded buildings, points immediately next to a gateway or missing calculations.
Technical details
Grid, formulas and diffraction model
The obstacle grid has a cell spacing of about 3 m. Roof surfaces are interpolated from triangles; unresolved edge cells conservatively receive the maximum roof height. Terrain is based on bilinearly interpolated DGM values using the DHHN2016 height reference. Radio paths are sampled in a local metric coordinate system at intervals of no more than about 1.5 m.
Free-space path loss = 32.44 + 20 · log₁₀(f in MHz) + 20 · log₁₀(d in km)
Received level = sensor EIRP + antenna gain − cable loss − free-space path loss − diffraction loss
Margin = received level − receiver sensitivity − planning margin − additional loss
Diffraction uses the dominant knife-edge parameter v and the single-edge formula J(v): for v ≤ −0.78, J = 0; otherwise J = 6.9 + 20 · log₁₀(√((v − 0.1)² + 1) + v − 0.1). Positive dominant obstacles are examined on both subpaths to a depth of three levels. An Earth-curvature correction uses an effective Earth-radius factor of 4/3.
This is a limited approximation based on the Deygout principle with the single-edge formula from ITU-R P.526, not a full implementation of the recommendation.
Where the model reaches its limits
Lateral diffraction, reflections, vegetation, interference, specific wall penetrations and actual antenna patterns are not included. Results may be too optimistic or too pessimistic. Sensitivity must match the gateway, spreading factor and bandwidth.
The prediction describes only the uplink. It does not indicate downlink capacity, delivery probability, regulatory compliance or range outside the village model. On-site measurements are needed for reliable decisions. Sites and settings are retained when you switch to these explanations in the same browser tab.
Antennas: a diagram is not a calculation profile
The reviewed data sheets include actual radiation patterns. For calculations, frequency, angular reference, mounting orientation and gain scale must also be unambiguously identified. VSWR or gain-versus-frequency curves alone do not describe directivity.
RAK 8 dBi
Horizontal and vertical patterns at 868 MHz are available. Their angular reference relative to the installed antenna and digitised curve values still need checking.
View the product in the iot-shop ↗RAK 5.8 dBi
Radiation patterns are available. The sheet names RAKARG13, while the diagrams name RAKARG14. No profile will be calculated from them until this is clarified.
View the product in the iot-shop ↗Fei Teng 6 dBi
A test report with E-/H-plane patterns and frequency data is available. The angular curves still need to be matched to the specific measurement setup and digitised with verification.
View the product in the iot-shop ↗The demo currently continues to use constant antenna gain. A separate selection of actual radiation profiles will follow only after the curve data have been checked; the product links do not change the calculation model.
Data and foundations
Official geodata and data sources · Calculation basics and original source →
The calculation uses the single-edge formula from P.526-16, Section 4.1. The model assumptions described here correspond to the SMARTinfeld demo.