3D Virtual Platform for Wireless Sensor Network Simulation – PERSEPTEUR

The PERSEPTEUR project has evolved from an alliance of laboratories recognized by the CNRS for their research on sensor networks and their interactions, physical and spatial modeling of the radio channel, and the cooperation with the company Virtualys specialized in 3D representations of urban structures. PERSEPTEUR aims to study and develop tools based on open urban descriptions and allowing the visualization of the characteristics of radio signals that cover a wide technological spectrum. In addition to spatial aspects, the project also aims to integrate temporal aspects that are based on mobility scenarios. This project intends to answer various questions such as the use of space and frequencies and its optimization. It also intends to give access to the information to the citizens (for instance on the existence and levels of radio signals) and to help deploying reliable monitoring services (like gas, light or other resources) by distributed sensors with a feedback to inhabitants.

Deployments of urban radio networks involve prior modeling and simulation. Design tools must meet three requirements: (1) be accurate, and quantify the model precision, (2) allow the representation of mobile nodes and dynamic environments and (3) capture the behavior of one or more networks at the scale of thousands of nodes.

Representing a dynamic environment is in itself a difficult task, with at the extreme the precision of the mobile movements and their radio connections. It is first necessary to represent a stable urban environment deterministically, and the fluctuation of the environment randomly (trees, barriers, climate).

The main challenge of this project is to find the right balance between different approaches and to control the errors induced by the statistical or deterministic missing information under an acceptable calculation time constraint. If this balance is obtained, it is clear that it makes possible investigations for developing reliable and optimized sensor networks in terms of communications, access to resources or routing. The gain of industrials is measured in terms of investigation rapidness and of reliability of the deployments.

The project is built around models for nodes’ positions (3D) and mobility, for environment evolutions, for the radio channel structure and for interference. The disposition of the nodes represents the strain of deterministic calculations. Statistical modeling of interference aims to simplify the calculations and quickly deal with a large number of nodes.

Simulations operate at multiple scales: buildings, streets, city. Calculations returns a map of the spectral occupancy, network and field levels visualization, or for a professional working in the deployment of a network, they evaluate a priori the interference levels and the reliability of links.

The work of this project can serve as a basis for the study and understanding of sensor networks, avoiding uncontrolled deployment. It promotes long-term research, in a realistic environment, about the maintenance, the performance and the energy consumption of nodes. It will also help citizens to visualize information about existing electromagnetic fields levels or to access to environment services. This work will also have beneficial effects in education. For example, complex propagation phenomena, interference or mobility will possibly be clearly illustrated.