Ageism
4.9. Bienestar psicológico y envejecimiento saludable
penetration loss.
4.5.3.3 Automatic SPM Calibration
The goal of this tool is to calibrate parameters and methods of the SPM formula in a simple and reproducible way.
Calibration is based on imported CW measurement data. It is the process of limiting the difference between predicted and measured values. For a complete description of the calibration procedure (including the very important prerequisite filtering work on the CW measurement points), please refer to the User Manual and the SPM Calibration Guide.
The following SPM formula parameters can be estimated:
• K1, K2, K3, K4, K5, K6 and K7
Figure 4.15Tx-Rx profile
Notes:
• To consider indoor losses in building only when using a deterministic clutter map (clutter height map), the 'Indoor Coverage' box must not be checked in predictions unless this loss will be counted twice inside buildings (on the entire reception clutter class and not only inside the building).
• Like for any Hata-based model, is, by default, limited to the computed free space loss value. It is also possible to avoid this option (option in the related scrolling menu of Configuration tab)
• Even with no clearance, the clutter height (extracted either from clutter class or clutter height folders) is never considered at the last profile point.
Lmodel
• Losses per clutter class (Kclutter must be user-defined)
• Effective antenna height method
• Diffraction method
Automatic model calibration provides a mathematical solution. The relevance of this mathematical solution with a physical and realistic solution must be determined before committing these results.
You must keep in mind that the model calibration and its result (standard deviation and root mean square) strongly depend on the CW measurement samples you use. A calibrated model must restore the behaviour of CW measurements depending on their configuration on a large scale, and not just totally coincide with a few number of CW measurements.
The calibrated model has to give correct results for every new CW measurement point in the same geographical zone, without having been calibrated on these new CW measurements.
4.5.3.3.1 General Algorithm
Propagation model calibration is a special case of the more general Least-Square problems, i.e. given a real m x n matrix A, and a real m-vector b, find a real n-vector x0 that minimises the Euclidean length of Ax - b.
Here,
m is the number of measurement points, n is the number of parameters to calibrate,
A is the values of parameter associated variables (log(d), log(heff), etc.) at each measurement point, and b is the vector of measurement values.
The vector x0 is the set of parameters found at the end of the calibration.
The theoretical mathematical solution of this problem was found by Gauss (around 1830). Further enhancements to the original method were proposed in the 60's in order to solve the numerical instability problem.
In 1974, Lawson & Hanson [2] proposed a theoretical solution of the least-square problem with general linear inequality constraints on the vector x0. Atoll implementation is based on this method, which is explained in detail in [1].
4.5.3.3.2 Sample Values for SPM Path Loss Formula Parameters
The following tables list some sample orders of magnitudes for the different parameters composing the Standard Propagation Model formula.
K1 depends on the frequency and the technology. Here are some sample values:
The above K1 values for WiMAX are extrapolated estimates for different frequency ranges. It is highly recommended to calibrate the SPM using measurement data collected on the field for WiMAX networks before using the SPM for
References:
[1] Björck A. “Numerical Methods for Least Square Problems”, SIAM, 1996.
[2] Lawson C.L., Hanson R.J. “Solving Least Squares Problems”, SIAM, 1974.
Minimum Typical Maximum
K1 Variable Variable Variable
K2 20 44.9 70
K3 -20 5.83 20
K4 0 0.5 0.8
K5 -10 -6.55 0
K6 -1 0 0
K7 -10 0 0
Project type Frequency (MHz)
K1GSM 900
935 12.5GSM 1800
1805 22GSM 1900
1930 23UMTS
2110 23.81xRTT
1900 23WiMAX
2300 24.7
2500 25.4
2700 26.1
3300 27.8
3500 28.3
predictions.
All K paramaters can be defined by the automatic calibration wizard. Since Kclutter is a constant, its value is strongly dependant on the values given to the losses per clutter classes. From experienced users, the typical losses (in dB) per clutter class are:
These values have to be entered only when considering statistical clutter class maps only.
4.5.3.4 Unmasked Path Loss Calculation
You can use the SPM to calculate unmasked path losses. Unmasked path losses are calculated by not taking into account the transmitter antenna patterns, i.e., the attenuation due to the transmitter antenna pattern is not included. Such path losses are useful when using path loss matrices calculated by Atoll with automatic optimisation tools.
The instance of the SPM available by default, under the Propagation Models folder in the Modules tab, has the following characteristics:
• Signature: {D5701837-B081-11D4-931D-00C04FA05664}
• Type: Atoll.StdPropagModel.1
You can access these parameters in the Propagation Models table by double-clicking the Propagation Models folder in the Modules tab.
To make the SPM calculate path losses excluding the antenna pattern attenuation, you have to change the type of the SPM to:
• Type: Atoll.StdPropagModelUnmasked.1
However, changing the type only does not invalidate the already calculated path loss matrices, because the signature of the propagation model is still the same. If you want Atoll to recognize that the SPM has changed, and to invalidate the path loss matrices calculated with this model, you have to change the signature of the model as well. The default signature for the SPM that calculates unmasked path loss matrices is:
• Signature: {EEE060E5-255C-4C1F-B36C-A80D3D972583}
The above signature is a default signature. Atoll automatically creates different signatures for different instances of the same propagation model. Therefore, it is possible to create different instances of the SPM, with different parameter settings, and create unmasked versions of these instances.
You can change the signature and type of the original instance of the SPM, but it is recommended to make a copy of the SPM in order not to lose the original SPM parameters. So, you will be able to keep different versions of the SPM, those that calculate path losses with antenna pattern attenuation, and others that calculate path losses without it.
The usual process flow of an ACP working on an Atoll document through the API would be to:
1. Backup the storage directory of path loss matrices.
2. Set a different storage directory for calculating and storing unmasked path loss matrices.
3. Select the SPM used, backup it’s signature, and change its signature and type as shown above.
4. Perform optimisation using the path loss matrices calculated by the unmasked version of the SPM.
5. Restore the type and the signature of the SPM.
6. Reset the path loss storage directory to the original one.
Dense urban
From 4 to 5Woodland
From 2 to 3Urban
0Suburban
From -5 to -3Industrial
From -5 to -3Open in urban
From -6 to -4Open
From -12 to -10Water
From -14 to -12Note:
• The Standard Propagation Model is deduced from the Hata formulae, valid in the case of an urban environment. The above values are consistent since they are normalized with respect to the urban clutter class (0 dB for urban clutter class). Positive values correspond to denser clutter classes and negative values to less dense clutter classes.
Notes:
• It is not possible to calibrate the unmasked version of the SPM using measurement data.
• You can also use Atoll.ini options, AngleCalculation = 2000 and AngleCalculation = 3000, for calculating unmasked path losses and angles of incidence, respectively. These options are only available for the propagation models available with Atoll by default. Please refer to the Administrator Manual for details.
4.5.4 WLL Propagation Model
4.5.4.1 WLL Path Loss Formula
Where is the free space loss calculated using the formula entered in the model properties, is the diffraction loss calculated using the 3-obstacle Deygout method, and is the diffraction multiplying factor defined in the model properties.
4.5.4.2 Calculations in Atoll
4.5.4.2.1 Free Space Loss
Please refer to the Appendices for further details about free space loss calculation.
4.5.4.2.2 Diffraction
Atoll calculates diffraction loss along the transmitter-receiver profile built from DTM and clutter maps. Therefore, losses due to clutter are taken into account in diffraction losses. Atoll takes clutter height information from the clutter heights file if available in the .atl document. Otherwise, it considers average clutter height specified for each clutter class in the clutter classes file description.
The Deygout construction (considering 3 obstacles) is used. This method is detailed in the Appendices. The final diffraction losses are determined by multiplying the diffraction losses calculated using the Deygout method by the Diffraction multiplying factor defined in the model properties.