|
Kategorie: Diplomové, bakalářské práce |
Tento dokument chci!
Tato práce zkoumá vhodnost a požadavky návrhu simulací pro simulátor NS-3 pro případ bezdrátových sítí používaných v měřící infrastruktuře společnosti Kamstrup. V práci jepopsán simulátor NS-3 a je vytvořena základní implementace dvou protokolů. WirelessM-Bus jako příklad jednosměrného protokolu pro zařízení napájené z baterií. Simulace Wireless M-Bus je porovnána s daty naměřenými v reálném systému. NS-3 poskytuje flexibilní prostředí pro vývoj simulací různých síťových protokolů, včetně těch určených pro sítě inteligentních měřidel.
3.
Figure 3. The number occurences on
the Y-axis represents the number concentrator-meter pairs that differed an
amount given the X-axis.
The maximum achievable hit rate this case, given the average transmission
period 3600/16 225.
The plots show that even the log-distance propagation model yields meaningful
results, even though they obviously not get close the simulation with known
average path losses. The other peak appears for high hit rates that are close the
maximum, creating the U-shaped plot.
28
.7 shows how many concentrator-meter pairs axis) were there for
a given range hit rates.
The Figure 3. The range centered zero difference the number of
concentrator-meter pairs that differed the least.7 and 3. There high number pairs that have very
low hit rates.
Figures 3.8 shows how the error estimation between the measured average
hit rate and the simulated ones was distributed.datasets differed because the system was not yet completely deployed, the mea-
surements were not complete, but the coordinates used for the simulation included
the locations that were planned for the future.5 Summary
Wireless M-Bus simple, one-way protocol, well-suited for simple battery-powered
devices. The left half the plot are pairs
that were “optimistic” the simulation, because the simulated hit rate was higher
than the measured one. The right half, the other hand, are the values that were
simulated with lower average hit rate. These are all the pairs that have very poor connection, but least
one packet got delivered due fluctuations the signal strength, they appeared
in the database. The model implemented this project coarse approximation it,
nevertheless closely predicted the general packet delivery rate simulation
based real data. The predictions could improved using more sophisticated propaga-
tion loss model, especially when more data about the installation the meters is
available, and using probabilistic model for dropping packets. The plot drawn using lines for clarity, however it
represents histogram with bins.
The model could further improved implementing model repeater
device.8 compare the hit rate values from the measurements with
the simulated hit rate using measured path loss and the simulated hit rate using
simulated path loss values
