Hi everybody After Philip Levis's comments on the strange fact that the LQI is better (in our experiments) for predicting distances, I was doubtful and began to wonder if there wasn't an error in our experiments. So I decided to (re-re-)check the code used in our experiments and I found a damn bug which caused the LQI and the RSSI to be switched in our log files :/
We ask everybody not to consider this article until we've corrected it. Sorry for this mistake. Van der Haegen Mathieu On 7/6/07, Philip Levis <[EMAIL PROTECTED]> wrote:
On Jul 6, 2007, at 4:09 AM, Van der Haegen Mathieu wrote: > I'd like to add that RSSI doesn't seem really correlated with > distance in the experiments we did with our sensors nodes in an > inside environnement with distances going from 1 to 10 meters. > > Lqi seem to be more correlated with distance in the same experiment > > I can send you our paper about this if you are interested > This seems strange, unless you were in a very short distance range; the CC2420's LQI is the soft chip decision coefficients, and so is directly a result of the SNR (and therefore the RSSI). Your assertion would only make sense if your links were all in the grey region of intermediate link quality, in which LQI varies. Phil
_______________________________________________ Tinyos-help mailing list [email protected] https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
