... | ... | @@ -17,7 +17,18 @@ I have plotted the delivery ratio, reception delay, and total transmissions for |
|
|
![reception_delay](/uploads/db60fd7965b2fd84899b5d840858ddd4/reception_delay.png)
|
|
|
![total_transmissions](/uploads/6912f6398ee1819cd7288f916c0ca9ba/total_transmissions.png)
|
|
|
|
|
|
I have done some statistics on the data sets but I am not sure how meaningful they are:
|
|
|
I have a couple hypotheses about what may be going on:
|
|
|
|
|
|
1. The later data-sets have a lower activity ratio `duty cycles with contact / (0.5*nodes*(nodes-1) * duty cycle count)`.
|
|
|
2. The earlier data-sets have significantly higher contacts than the later sets.
|
|
|
3. The later data-sets may have issues relating to the Android task scheduler not letting Bluetooth scan exactly every 5 minutes which causes a shifted duty cycle.
|
|
|
|
|
|
Computing the average degree of each of each of the nodes in the entire network we see that it is corelated with the delivery ratio for both algorithms
|
|
|
|
|
|
![average degree](/uploads/6a113c000da4c5b9e34be71de69889ed/chart.png)
|
|
|
![Delivery ratio](/uploads/af754082e2010ce6956b36958ad3302d/delivery_ratio.png)
|
|
|
|
|
|
I have done some other statistics on the data sets but I am not sure how meaningful they are:
|
|
|
|
|
|
**shed/data/pp_shed1.csv**
|
|
|
|
... | ... | |