Is it time diversity, if I change the advertising interval of the beacon?
Not sure I understand what you mean by “time diversity” here … Could I ask you to expand a little bit? (:
From papers that I have read (the papers talk about diversity technique to combat fading), there are some diversity technique that we can use to combat signal fading that is caused by multi path propagation. One of the technique is time diversity, the general explanation of time diversity is “The information signal is transmitted repeatedly in time at regularly intervals. The separation between the transmit times should be greater than the coherence time, Tc. The time interval depends on the fading rate, and increases with the decrease in the rate of fading.”
Since we can adjust the advertising interval of estimote beacon, I assume that by adjusting the advertising interval of estimote beacon then I already have done the time diversity technique.
So my question, is my assumption right?
Thank you
Ah, of course, silly me! I know about this technique, just didn’t make the connection.
Unfortunately, it’s a bit more complicated than that—the time differences are super tiny and require extreme precision to get right. Our current beacons don’t have adequate hardware support to implement this.
I assume that you’re looking into this b/c you require very precise distance measurements, is that right? What’s your project/app/use-case about?
Yes you’re right, so my thesis is about how to increase accuracy of distance measurement between the beacon and my phone by doing time diversity technique(unfortunately, after I read your explanation, it seems like decreasing advertising interval is not time diversity technique at all). So I measure the accuracy from four different intervals: 950ms, 800ms, 500ms and 300ms. I found that, yes it is true, by decreasing the advertising interval, the accuracy is increasing.
The accuracy (in %) for each interval are: 72%, 74%, 79%, 88%.
So now I have a different question. Would you like to explain to me scientifically, what is the correlation between advertising interval and accuracy of of distance. Why if the advertising interval is decreased then the accuracy of distance measurement will increase?
The code/formula that is used to compute distance proximity is:
public static double computeAccuracy(Beacon beacon) {
if(beacon.getRssi() == 0) {
return -1.0D;
} else {
double ratio = (double)beacon.getRssi() / (double)beacon.getMeasuredPower();
double rssiCorrection = 0.96D + Math.pow((double)Math.abs(beacon.getRssi()), 3.0D) % 10.0D / 150.0D;
return ratio <= 1.0D?Math.pow(ratio, 9.98D) * rssiCorrection:(0.103D + 0.89978D * Math.pow(ratio, 7.71D)) * rssiCorrection;
}
}
I get that code from estimote-sdk-preview.jar, and would you like to explain where did the constant (0.96, 3 etc) come from?
Thank you, sorry if I ask you too many questions.
All the constants are a result of our engineers doing some math to find a function that best approximates changes to RSSI when distance changes (:
As for why decreasing the advertising interval improves accuracy: we’re doing some averaging/smoothing out the RSSI in our SDK, and if the SDK receives more “pings” from the beacon, it’s averaging algorithms have more data to work with, and can produce better results.