Radio Planning

In the Netherlands there are 13 radio channels allowed in the 2.4 GHz frequency band to be used for radio local area networks. Due to the DSS technology a used channel does not occupy a single discrete frequency but is a distribution of power in a 22 MHz wide frequency interval. In figure 2 it is shown that there are 3 completely separate channels available. Combining this knowledge with the topology and the goal of providing coverage in a fairly big area poses a challenging problem [Beckmann]. This problem is solved by careful selection of the channels to be used. For example, at the cost of a small increase in noise but no decrease in bandwidth, it is possible to use 4 channels on the same site. (e.g. channels 1,5,9 and 13). Of course the channel use is very much dependent on the local situation at the radio-level. Other measures to prevent cross-interference of different radio-links include the use of (directional) antennas and the location and polarization of the antennas.

Figure 2: The available channels in the Netherlands.

In the Netherlands the 2.4 GHz frequency band is reserved for a number of licensed and unlicensed applications ranging from microwaves to video-links, vehicle identification systems and radio amateur (ham) use. The use of WiFi in this so-called ISM (Industrial, Scientific, Medical) band is allowed unlicensed, providing the above mentioned channel restrictions are observed and the effective output power of the antenna (EIRP) does not exceed 20 dBm (= 100 mW).

Providing coverage in the target area requires careful radio planning. Often chaotic behaviour is seen (i.e. small changes in initial conditions can turn out to be critical). Techniques such as hexagon planning, as used in the cellular telephone world, are used to optimize the coverage and spectrum (channel) efficiency. A homogeneous network is created by placing every site on a predetermined grid point such that every cell covers half of the inter cell distance. The physical cell boundaries are determined by the local situation and the antennas used. The cell boundaries are not symmetrical for up and down link due to local and receiver noise. This also needs to be taken in account to avoid areas without coverage (i.e. downlink possible, but no uplink possible). Using commercial Hata-Okumura [Ho] model based, radio-planning software we can simulate the propagation of the signal and optimize the location of the different sites. These advanced planning tools (e.g. CellCad (www.lcc.com) or PathLoss (www.pathloss.com)) are not comparable to those available in Open Source. Alternatives such as Radio Mobile (www.cplus.org/rmw) lack important data on the environment such as building characteristics. As most of the commercial tools require a large computing infrastructure and various subscriptions to maps and other GIS data, like building heights and absorbency or reflection characteristics, it is not feasible to actually run the simulations ourselves. A sponsor has access to this software (CellCad) and runs the required simulations.

Nevertheless, a site survey is always needed to actually measure the noise generated by other radio sources on-site and check the signal strength of the already running nodes. Due to the high absorbency of the radio signals by (for example) trees in the line of sight, local field measurements are essential in planning a node. A site survey is done using the Kismet [Kismet] or dstumbler [Dstumbler] software running on a Linux or FreeBSD laptop with a small panel or directional antenna and wireless network interface.

As a subproject within the community a compact automatic survey tool is currently under design. This tool will collect signal strength and noise figures from existing Wireless Leiden nodes as well as signals from access points or other devices in the 2.4 GHz band. The hardware device based on an embedded system connected to a GPS receiver can be carried through the city on various utility vehicles to effectively cover the complete area. Once completed, the software environment will be comparable to the software setup of a node with some extra tools to log the data. When available, the logging data will then be plotted as an overlay on the simulation maps and made available to the users wanting to connect to the network as a guideline for aiming their antennas.