Hi, in GSM localization application, can anyone tell me the relationship between the signal noise measurement and the variation in signal distance.
My research field is Localization Based Services in GSM system using fingerprinting method. Which is based on the signal distance computation.
Dear Belaabed,
The signal level at any point around a base station depends on the effective isotropic radiated power , the path loss and the receive antenna gain. These are the signal budget parameter in a radio channel. The effective isotropic power depends on the transmitted power and the transmit antenna gain. The path loss depends on the propagation distance from the transmitter to the point of reception and the wireless channel model. The wireless channel model may be Rician or Rayleigh depending on multi path effects present in the channel. These channel models are well described in the literature. The mulipath effect leads to statistical variations in the signal amplitude called fading. This concerns the received signal at the input of he transmitter. The receiver itself issues noise which is added to the received signal causing the contamination of the signal and thereby leading to errors in the correct signal detection specially when the the received signal is relatively weak. It is so that the noise is mostly generated in the receiver and therefore the channel is modeled by adding a white noise source and the channel model is called an additive white noise. The signal power to noise power ratio determines the bit error rate for a given modulation technique.
wish you success
Thinks Mr Abdelhalim for your valuable informations.
I want to give you some details to my problematic in this answer.
I synthetize the radio-map for database correlation method using Okumura-Hata model (For cellular network positioning system). At the receivers, I add an additive white noise to the measurements. During the localization phase, the location system computs the signal distances between the fingerprints and the online measurements, and by using the matching algorithm the system votes the closest fingerprints to the real time measurements.
To introduce an optimal number of the selected fingerprints in this phase, I want to add a threshold to this matching algorithm to comput the best selected fingerprints, which is (threshold) in my mind related to the noise.
Thank at all
Localization of a transmitter, just on the basis of a standard equation is fraught with many a possible loopholes. The relative measurement method or shadowing model will give better results.
Even then, a factor of uncertainty must be added to make the results realistic. This variation can be Gaussian or some other pattern. The variation has to be considered to take care of fading.
The best way to work out the variation is to take a number of readings and finding the variations from the calculated readings. RF communication and prediction is a pretty challenging task.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing