In: Products, Resources 15 May 2017 Tags: , , , , , , , ,

Efficiently planning mesh networks with HTZ (Part 1 – Digital cartography).

This post is intended to help a radio-planner, technical director, project manager, or consultant to be more aware of the important goals to pursue when planning large scale mesh networks in urban environments. It proposes innovative ways to accurately manage large areas of interest, using cartographic data with mixed high and medium resolutions.


The radio-planning of mesh networks can be divided into three main topics: – Dimensioning the mesh node distribution in order to achieve the requirement of coverage of the end user. – Analyzing the linkage of the Mesh Nodes, in order to optimize the dynamic routing and therefore ensuring demand throughput. – Backhauling the gateways (Microwave links…).

  • Required components
    Cartographic data Mesh network planning can be achieved by using different kinds of digital cartography. The choice depends on the data already available, the budget to be spent, the time available for the planning and the accuracy to reach. The user usually is able make the choice between two types of datasets: – Medium Resolution cartography – High Resolution cartography giving exact locations and heights of the buildings for a given area of interest. However, both datasets have their pros and cons because the areas to treat during mesh planning are large. A third dataset choice known as a hybrid dataset, combines the advantages of both other types of cartography and will be highlighted throughout this document.


Medium Resolution for large areas (From 10 to 30 m resolution)
A typical medium cartographic dataset contains the following layers:

– A Digital Terrain Model

– A clutter file, provides a description of the ground occupancy as major aggregates (urban, dense urban…)

– Topographic or Aerial maps (Online or Offline)


Adapting the medium cartography to mesh planning Standard Medium Resolution datasets usually do not feature the roads, because their width is usually smaller than the resolution of the dataset itself. Roads and streets are a crucial component for mesh radio-planning.  The mesh nodes are usually installed in the streets in order to take advantage of the canyoning effect. If the roads and the streets are not available, HTZ features a drawing interface allowing the radioplanners to add the streets in clutter file by importing the street network from a GIS database (OSM, National Geoportal, GIS collections…)


Pros and Cons

The Medium Resolution dataset for mesh planning offers several advantages:

– It allows the treatment of very large areas: the planning of an entire large city can easily be achieved

– The availability of this data is quite good, usually for a reasonable cost

– The resolution of the data allows very fast computations

– Prediction values can be compared with mobile measurement data

However, the radio-planner has to keep in mind that a reduced resolution for the cartography usually generates a reduction in the planning accuracy. Because of the sensitivity to the building environment of the mesh frequency (usually in the ISM bands), using a coarse cartography will generate a coarse planning result that might not reflect the planning accuracy that is targeted (especially for the “hot-zone(s)). Also, the cartography must be manually treated in order to “dig” the streets in the dataset (if the data is not already available in GIS format), in order to provide the ability to simulate the canyoning effect between the mesh nodes.

Availilbity: Worldwide (<=30 m)

Planning accuracy (standard deviation):

– HTZ (Deygout94 propagation model): <=5 dB

– Empirical propagation models (3GPP / Seamcat / Cost): <=8 dB

– Half determinitistic models (ITU / FCC): <= 7 dB

– Coordination models (ITU / HCM): <= 10 dB

Cost-effectivness (Dataset building workload): < 2 days (depends on surface and resolution)

Vintage dependency (Free Clutter data): between 5 and 10 years old


High Resolution (from 1 m to 5 m)

A typical high resolution cartographic dataset contains the following layers:

– A high resolution Digital Terrain Model (featuring the bridges for instance)

– A building height file, for the canyoning effect and for building diffusion loss

– A clutter file that describes the vegetation and the different “types” of buildings (concrete, glass…)

– Ortho-photos or Aerials


Pros and Cons

High Resolution cartographic data provides precise building location and height for a specified area of interest. It is the optimal product to achieve excellent planning accuracy in outdoor (canyoning effect according to the exact shape of the buildings) and also in indoor (signal diffusion according to the building type) environments, even though it might require longer computation time.

– Prediction values cannot be compared with mobile measurement data (only fixed measurement data)

However, this type of data remains expensive. It is therefore advised to be used on small areas, and centred, for instance, on “hot-zone(s)”. The coverage calculated also depends on the building data available: the vintage of the production source is also quite important.


Availilbity: low but can be built on demand (around 150 EUR/km²). Some geographical institutes offer HR data for free or at a very attractive price (< 20 EUR/km²)

Planning accuracy (standard deviation):

– HTZ (Deygout94 propagation model): <=4 dB

– Empirical propagation models (3GPP / Seamcat / Cost): <=8 dB

– Half determinitistic models (ITU / FCC): <= 6 dB

– Coordination models (ITU / HCM): <= 10 dB


Hybrid cartographic dataset

Another option in preparing a workspace for the planning of a WiFi mesh network would be a hybrid cartographic dataset. The hybrid dataset uses medium resolution data for the areas where a high degree of simulation accuracy is not critical. The hybrid dataset uses high resolution data for “hotspots” within the user’s planning workspace to optimize simulation accuracy for critical areas. High resolution data includes exact building dimensions centred over these “hotspots.”

The hybrid dataset contains both medium and high resolution layers of cartographic information:

– The DTM covering the entire Area Of Interest, with more details in the hot-zone

– A Clutter file mixing the Medium Resolution data and some High Resolution data

– A building height layer covering the High Resolution areas only (the urban heights in the Medium Resolution are managed in the clutter file)

– Imagery (Online or Offline)

– Vector layer (3D building polygons, transportation…)



Pros and Cons

The hybrid dataset combines the assets of the two datasets previously defined. For most areas, coarse radioplanning is performed with both time and cost effectiveness, whereas the “hot-zones” can be locally analyzed with high planning accuracy. The clutter file requires a careful configuration, because it mixes medium and high resolution data:

– The medium resolution clutter codes are configured with average heights, that will be used by the NLOS (diffraction) engine of HTZ when the final receiver is located inside these clutter codes (Suburban for instance)

– The high resolution clutter codes are configured with no heights (as they are defined in the building height file), but with a diffusion loss coefficient in dB/km. This will be used when the final receiver will be located in a building in a “hot-zone”.




High Resolution Terrain & Clutter Datasets: Why Lidar?
There are myriad methods, techniques and technologies for obtaining elevation and earth cover information through propagated signals. Those technologies may be based on sound, radio and light and also vary in resolution, difficulty, expense and process. Overall, most of these sensing technologies are based on the time delay of a reflected or scattered signal, though traditional passive sensors can also be used and rely on natural radiation.
Lidar systems illuminate a target with lasers, then receive and process the reflected and/or scattered signal. Modern Lidar systems are compact, precise, and efficient and provide many advantages over traditional photo-based techniques. They allow for sub-1 meter data collection and improvements in post-processing aid in the ease of use of the data. Most post-processed Lidar data is classified by return number and category, further shortening the conversion process from raw data to datasets that are usable in RF propagation tools.
Another advantage of Lidar data collection is that the data may be collected both day and night unlike traditional methods which require collection during daylight. Lidar not only offers high accuracy, but allows for the collection of elevation information in areas of dense vegetation. Since a Lidar pulse can have multiple reflections, it will reveal both surface elevation and terrain elevation at any point. Most other collection techniques only gather information about surface heights. Furthermore, modern Lidar data collection systems are compact and are easily mounted onto light aircraft for data collection over large areas.

Get Lidar Data
Lidar, or light detection and ranging, can be used to quantize terrain, ground clutter and ground occupancy. Airborne Lidar systems are typically used for the purposes of scanning large areas and are composed of a laser and a rotating mirror that is used to sweep the area of interest. The airborne Lidar system then acquires data points by bouncing a laser signal off of the earth, buildings and vegetation. As the airplane flies, the Lidar system quantizes the terrain and ground clutter below in a zigzag pattern, as pictured below.


The acquired data points are reflections of the laser signal from obstacles in its path, and there may often be multiple reflections of a single emitted signal. One reflection may be produced by buildings, the ground and other solid objects. Trees and vegetation may produce several reflections as the laser signal propagates through the leaves and reflects off of branches and ultimately the ground. Thus, it is common to have multiple returns for a given transmission.
To compute the distances between the airborne Lidar sensor and the reflection point and thus the elevation of the reflection point, calculations are run using the elapsed time and the speed of light. This data is correlated with the GPS positioning of the aircraft along with inertia sensors or gyroscopes to accurately create the environment of three dimensional points that are a Lidar dataset.

Lidar Data Format
Lidar information is typically obtained and stored in ASPRS LAS format. Not only does the LAS format contain information about surface heights, but it also provides header information that contains technical information such as return number, classification, and scan angle, among others. The user may then employ or develop software that sorts through the Lidar data (often in the realm of Gigabytes of data) to create the desired datasets based on required criteria, such as return number or classification.
Once the Lidar data has been sorted as desired, the data can then be manipulated as necessary. In most cases, this involves interpolation of the dataset to create a continuous model without non-return pixels. Lidar data is, by nature, discontinuous since individual measurements are based on specific geographic points. These points are then stored in the LAS file. In essence, a LAS file is a list of measurement information per geographic point. In Figure 2 the image on the left shows a dataset where only the first return is shown.  Any areas in black are points where Lidar data was not obtained during the measurement campaign.  In the image on the right, the dataset is interpolated into a continuous dataset that can be used in radio frequency planning and analysis software as a digital surface model.


Lidar Dataset Preparation for RF Analysis

The images in Figure 3 and descriptions that follow show how ground occupancy and clutter are generated using processed Lidar data:
Aerial Image: The first image is an aerial photo of the area of interest, showing the presence of roads, vegetation, buildings and unoccupied ground.
Bare Earth Image: The second image is a bare earth model that was extracted from the ground points of a Lidar dataset and then interpolated into a smooth, continuous dataset. The bare earth model is also known as the DTM, or digital terrain model, since it contains only ground elevations. The DTM is derived from a Lidar dataset by removing all points but those classified as ground points and then interpolating the data to create a continuous dataset.


First Return Image: The third image is a first return model that was extracted and then interpolated into a smooth, continuous dataset. The first return model may also be called a DSM, or digital surface model, since the heights and elevations it contains are the maximum elevations for the ground and any ground clutter at each point. The DSM is derived from a Lidar dataset by removing all returns but the first return and then interpolating the data to create a continuous dataset.
Ground Occupancy Image: The final image is a clutter dataset, obtained by subtracting the heights in the digital terrain model from the heights in the digital surface model, leaving only buildings, vegetation and any other ground clutter within the file.

Areas in black represent a lack of clutter, where the digital elevation model and the digital surface model are equivalent.


RF Analysis Using Lidar Data

Once the Lidar data is processed and converted to the appropriate formats, the user may load the datasets into HTZ to run simulations.

The high resolution Lidar dataset provides for highly precise modeling with sharp blockages.


The above image, is a three dimensional visualization of how the sample RF signal is incident upon the exterior of a large building.  The color variation across the building’s facade and stepped roof represents varying power received levels of the RF signal.

The Case for Lidar Datasets

While traditional terrain and surface cover collection techniques and technologies yield resolutions that often range in the tens of meters and can be as accurate as 3 to 5 meters, Lidar allows for sub-1 meter data collection. As a highly precise and detailed terrain and clutter format, Lidar data is well suited for high-resolution RF analysis. Not only is the user presented with unparalleled accuracy for propagation over bare terrain, but also for propagation over and through ground clutter.

In: Products, Resources 13 May 2017 Tags: , , , , ,

MF Groundwave Propagation Modeling for Maritime Networks
Introduction to modeling MF band propagation (3 kHz – 30 MHz) for Maritime Networks with HTZ
For the past seventeen years ATDI has been integrating and developing software for modeling anomalous radio wave propagation for the purposes of RF network design. This includes propagation phenomenon such as ducting, troposcatter and their applications over terrain and water.
Over the past five years, ATDI has dedicated significant resources into investigating how to model the propagation characteristics of frequencies below the VHF band. There are many applications to these frequencies including but not limited to:

  • Aeronautical Navigational Aids
  • Automatic Link Establishment for Intelligence gathering
  • Emergency communications for Maritime Networks


ATDI has developed several specific features into its product line for modeling a variety of below VHF band propagation for each of these applications, this document will be the first in a three-part series highlighting how ATDI’s flagship RF network design tools model Maritime Communications.
This first document will focus on modeling MF Groundwave propagation from ship to shore along coastlines. This paper will focus on developments in the areas of cartographic map data preparation, integration of propagation standards and calibration information and custom reporting options available to users of HTZ for the purposes of modeling Maritime Networks.


Preparation of a Conductivity/Permittivity Map from the ITU IDWM database:

In the case of MF propagation, terrain obstruction information provided by the classic Digital Terrain Model used with most RF network modeling packages is of greatly diminished importance. More important, are the electromagnetic properties of the terrain in particular the Conductivity and Permittivity of the ground.
These types of maps are usually available from the local national spectrum authority. ATDI’s GIS management tool, ICS map server tool can create these maps from any type of source (digitized map, vector map, etc.) in a format compatible for RF analysis with HTZ.
ATDI cartographic services can also provide this type of information for any country in the world using the ITU Digital World Map (IDWM) database as a global source of conductivity and permittivity data in all varying regions. Note, that this is the same source for the conductivity map in the FCC 47 CFR 73.190.

The map above is provided in the form of HTZ’s classic clutter layer. Since the clutter layer can serve as a generic skin or blanket of morphological information layered over the terrain model, and can contain user defined propagation characteristics per clutter class/code, this layer was perfect to reuse as a conductivity map layer.
The units of each region of conductivity are in milli-Siemens/meter (mS/m) and can be configured as labels of each clutter class/code to give the map distinction in the HTZ interface.
In order to properly model the radio wave propagation of MF signals, ATDI has also integrated the latest ITU recommendations specific to MF Groundwave propagation: ITU-R P.368-9 and ITU-R M.1467-1.
calculation feature used to generate the field strength received predictions for each pixel on the map is based on the integration of ITU-R P.368-9 into HTZ’s propagation engine.
The ITU-R P.368-9 model depends on the input of conductivity and permittivity data which is provided by the ITU maps described previously.
These values provide the ITU-R P.368 Groundwave model with the appropriate attenuation information to model MF propagation over land and sea allowing HTZ to generate MF Groundwave coverage plots.
In order to make sure that the receive sensitivity of each radio network element is configured appropriately, with respect to their immediate environmental conditions and time of year,
HTZ has also integrated a NOISDAT calculator derived from ITU-R M.1467-1.


The NOISEDAT Calculator takes into consideration the operating frequency, bandwidth, signal-to-noise ratio, 90% fade margin and estimated radiated power as well as specifications of the receiver environment and season to model the variability in Noise contribution to radio propagation in the MF band.
Essentially, the NOISEDAT calculator serves as a reference to model the expected Noise Rise and respective threshold degradation at a given site of interest.
ATDI has even integrated consideration for A2 sea region in order to generate an output based on ITU-R M.1467-1 NOISEDAT calculation to give the predicted receive sensitivity in dBm and dB-V/m as well as range in nautical miles and kilometers. This information is used to calibrate HTZ’s propagation engine appropriately for ship to shore (reverse coverage) calculations.


Reporting options specific to modeling Maritime Networks
ATDI tools also includes reporting features specific to modeling Maritime Communications including the ability to generate nautical mile boundaries from the coastline or from the locations of the shore stations:

ATDI continues to refine its modeling processes for MF Groundwave propagation studies in response to emerging requirements from the spectrum authorities of Coast Guards and Naval agencies all over the world.
ATDI’s strong association with the ITU, and expertise in integrating ITU recommendations into its product line allow ATDI to be the world leader in translating complex propagation phenomenon to simple, intuitive graphics that can be understood by the various policy makers and stake holders involved utilizing and managing a country’s spectral resources.
In upcoming parts of this series on modeling Maritime Communications, we will focus on newly developed features for generating probability of coverage per season and frequency for HF Skywave propagation as well as modeling HF antennas and ultimately VHF coverage and traffic analysis for Maritime Communications.


In: Services 15 Nov 2016 Tags: , , , , ,

The next generation of mobile communication, 5G, is now this side of the horizon but its form remains indistinct and shrouded in mist.


Despite the enormous research efforts into 5G mobile technologies, there remains uncertainty on what services will drive the deployment of 5G networks, and which of these services will be delivered by mm-waves.
ATDI technical director Nick Kirkman comments: “A consensus is emerging that important characteristics will include perceived ubiquity of service, very low-latency (for virtual reality, real-time control, etc.) and an order of magnitude increase in supported bit-rates.  While ‘traditional’ cellular frequencies below 6 GHz will continue to be critical to such networks, it is also anticipated that millimetre-wave frequencies will be exploited to allow high data rates over short distances.”
A number of frequency bands between 25 GHz and 86 GHz are currently being studied under the auspices of the International Telecommunication Union. Industry collaboration projects such as mmMagic (Horizon 2020) are looking to characterise radiowave propagation at frequencies above 6 GHz. However, the development of site-specific models for planning at these frequencies has received little attention.
Paul Grant, ATDI’s operations director, notes: “At these high frequencies, coverage planning techniques will be radically different from those used today. Diffraction losses are so high that there is an almost binary switch between ‘served’ and ‘unserved’ areas as shadowing by environmental clutter interrupts the line-of-sight between base station and user.
“It is likely that small cells will make extensive use of beamforming and MIMO techniques, required because of the high path losses and made feasible by the small antenna sizes. This implies that a useful prediction model will need to take account of scattered and reflected energy; this provides a challenge as it is a computationally more intensive task than the direct-path predictions generally used in operational planning models. Such predictions are also likely to require rather detailed data concerning building facades. The proposed study will be a valuable opportunity to understand the improvement such data can bring to coverage and capacity modelling.”
Nick adds: “The current trend in the mobile industry is towards self-optimising networks but this does not imply that there is no role for planning. With the cost of access to sites for base stations rising rapidly, it is an urgent matter to ensure that the best use is made of physical resources. The output of this project will help to minimise 5G network deployment cost and increase efficiency.”
The planning techniques being developed by ATDI are more sophisticated than anything that has been publicly reported for outdoor deployment of 5G mm-wave networks.
“We are aiming to provide insight into the viability of use cases for 5G at mm-waves,” Paul says. “One typical area would be the deployment density needed for base stations. This will be valuable both for government in developing its policy for 5G and industry in developing business cases for 5G.”