Approximating Radio Maps: an experimental study

This web page covers the Radio Map project,

A new algorithm for approximating Radio Maps over large (rural) terrains is suggested implemented and tested.

 Goal, Motivation, Radar-Like Algorithm, Links 

 

The figure above shows Radio Map Approximations as computed by the various methods: 5000 samples each (left to right):

(a) Grid,  (b) Random, (c) Terrain simplification, (d) Fixed radar, (e) sensitive radar.

 The corresponding extrapolation is given below each sample-set. The antenna is in the center and the brightness

 represents the strength of the signal. Observe that sensitive radar (e) has the best edge detection.

 

 

Goal   

Given a terrain T and an antenna A located on it, associate the signal strength of each point  on T as received from A -- approximate the Radio Map of A over T.

 Back to Top Page

 

 Motivation   

The rapid development in wireless communication technology has dramatically reduced costs and increased the number of users. This development has increased the demand for wireless communication while supporting better throughput and quality of service. To withstand the increasing demand, more antennas should be installed on urban and rural regions. However, the cost of installing antennas accounts for a significant fraction of the wireless communication hardware system. For that reason, communication companies are interested in installing a minimal number of antennas that provide the required throughput and quality of service. Solving such optimization problems, requires efficient methods to approximate a Radio Map

In addition, recently new wireless broadband technologies (namely $WiMAX$ or 802.16) can support 50-kilometer connections and more with bandwidth that exceeds current DSL and cable broadband capabilities. Meaning, single antenna can cover large area of sparse population without wires stretched into homes by local phone and cable companies. 
This ability leads to lower prices and faster consumer adoption.
Many communication companies that deals with design or optimizing the wireless network layout uses optimization tools  to optimize such antenna location. These optimization tools approximate many  Radio Maps which increase the need to efficient Radio Map approximation methods.

 Back to Top Page

 

 Radar-Like Algorithm (RLA)

We have developed a novel approach to approximate Radio Maps over a terrain. Our approach simulates a radar scanning scheme using an efficient pipeline technique to compute sample-points and signal strength along a radar ray that represents a cross-section of the terrain. 

The main idea of the pipeline method is based on computing several sample points along a single cross-section.  For a given transmitter A and receiver B, instead of just computing the signal strength at B - coming from A, with just a minor overhead, the signal strength at other points along the cross-section AB could be computed. The method is based on efficient 2D-terrain simplification heuristics.

Note: in order to predict the signal strength at a point, we used the following theoretical models: Single knife edge (Bullington) and Multiple knife edge (Epstein and Peterson, Deygout) propagation models.

 Back to Top Page

 

Links 

For more information regarding the Radio Map project, email me: benmo (at) yosh.ac.il

 Back to Top Page