Physical-statistical models for mobile satellite communication systems below 10 GHz (PhyStat)


The objective of the study was to develop a software tool implementing a set of channel propagation models that combine deterministic and statistical approaches including physical principles (diffraction, specular or diffuse reflection, transmission) together with stochastic parameters. The interest of these approaches is to be able to extend the validity of the models to different scenarios and environments. The models and associated software tool shall cover land-mobile satellite, aeronautical and maritime environments. They are intended to be used for system design and/or air-interface performance optimisation for frequencies below 10 GHz. Therefore they must provide both time-series representative of the channel and statistical distributions of the channel characteristics than can be used for systems design.


Various methods and models have been proposed in the past to describe the fading effects in the LMS propagation channel at large and small scales. They can be empirical, statistical or analytical/physical. Physical or site-specific models can provide precise and accurate results but are unsuited to our problematic since their computation time would be far too important in a cellular environment. On the other hand, empirical models, based on measurements campaign, are a fast way to provide accurate statistical results. However their results are strongly bound to the environments of the measurements and the frequency.
Therefore a compromise between these two approaches, known as physical/statistical model, has been proposed. The philosophy of this approach is to apply more or less simplified physical propagation models to urban or suburban scenarios drawn from statistical distributions of the characteristics of a given environment such as building height distribution or width of the streets, etc... This kind of models combines the statistical accuracy, ease-of-use and low computational requirements of empirical models, with the physical insights of deterministic models and the possibility to estimate the model for a lot wider range of elevation and azimuth situations.


The main benefits of the project are the development of advanced validated physical-statistical propagation models to be used for systems studies. Specific environments such as maritime, aeronautical and urban have been considered.
In this study, two types of achievements have been reached:
  • From a scientific point of view, several physical-statistical propagation channel models have been developed or improved:
    • The aeronautical model based on a GSCM.
    • The maritime model based on similar GSCM principle as the aeronautical one and implementing a model for the movement of the ship and sea waves.
    • The land propagation channel model for pedestrian/vehicular in urban or tree-sided road environment has been built, composed of both updated and entirely new elementary models:
-UTD is used for building and other obstacles masking (poles, lampposts, vehicles, tree trunk masking …), attenuation considered through ray-tracing and attenuation coefficient. Scattering by trees considered by a simple GSCM model with pre-calculated parameters.
-The new 3CM model has been integrated for building façade scattering, whereas a GSCM approach in the UCL model is an alternative for the same effect.
    • For terrestrial components, a CGC module based on an adaptation of the Winner model.
    • Standard antennas, and modified pattern for installed antennas on platform can be considered.
  • From a scientific point of view, several physical-statistical propagation channel models have been developed or improved:
    • All channel models, with a simple GUI allowing to describe the simulated environment and RF link (including antenna patterns) and to define all the model configuration parameters. .
    • A large set of Matlab post-processing modules for analysing CIRs or time-series from the main modules (e.g.  diversity schemes or interfacing external modules to analyse system issues).
    • Several visualisation Matlab modules
The main interests of the physical-statistical approaches allow to study either typical generic environment, or specifically built environment (to compare with measurements for example); and allows extending the validity of the models to different scenarios, for instance for frequency range.


The PHYSTAT simulator allows the user to select one of the following simulation modules:
  • aeronautical model,
  • maritime model,
  • vehicular and pedestrian model.

All those simulation modules can be started either in the PHYSTAT user interface or directly in a batch mode. The main output for all the simulation modules is a channel impulse response saved in a (.mat) data file. Several post processing and visualization modules can be applied to this channel impulse response. The implementation details on the PHYSTAT software are described in the Software Simulator User Manual.

click for larger image

Regardless to the simulation model selected by the user (aeronautical, maritime, vehicular and pedestrian), the primary output is the channel impulse response at the receiver side which contains amplitude, delay, angle of arrival and Doppler information. If needed, specific antenna patterns can be applied.


The logic of the study was based :
  • first on a review of existing mobile (satellite and terrestrial) propagation models available in literature that uses combined physical-statistical approaches, and a review of the datasets available for parameterising and validating the models
  • second the design of the global software simulator which includes different models or approaches
  • third the adaptation of existing models and the development of new models which were included in the simulator
  • fourth the integration of all these models into the simulator, and its complete implementation and validation
  •  fifth the use of the simulator to perform several case studies, and the reporting of the results

click for larger image

Current status

COMPLETED. The project has ended on October 15th, 2013.


Status date

Monday, January 20, 2014 - 15:38