Ted brought over an app that runs on three interacting motes. A sniffer picks up which of the motes received a message from another mote and the strength of the message received. The first test we want to perform is a compatibility test among the motes, to see if some just don’t play well with others.
Details below.
1) Begin with six motes with battery packs. Number them 2-7 (mote 1 is for synching)
2) Confirm that channel 25 is still quiet in the lab
3) Set the relative strength of all the motes to power level 3, which should work reasonably well for distances of about 1 meter.
4) With the first 3 motes (2-4) in consistent relative positions, run them for 2 minutes
5) Record the number of successful communications among each pair and the average and std dev. of the signal strengths.
6) Switch out the motes and repeat six times in the following order:
2 3 4
2 5 6
2 3 7
3 4 6
3 5 7
4 5 7
4 6 7
Your data table should look like this:
snd rcv trial (# successes/# attempts) (avg. strength) (std dev strength)
1 2 1
2 1 1
1 3 1
3 1 1
2 3 1
3 2 1
2 5 2
5 2 2
2 6 2
6 2 2
5 6 2
6 5 2
2 3 3
etc.
Display the results with the mote numbers in a circle and the connections between them as directed lines between the motes. Write the three values along each of the arcs. Some pairs will have 2 sets of data for each direction, since the same motes appear in two separate trials.
Analyze the data to determine if there are some motes that don’t communicate well with any of the others or if there are separate sub-groups of motes that communicate only with one another.
Note that this type of test grows quickly with the number of motes involved. The pattern is the number of connections for n motes is (n-1) + (n-2) + (n-3) + … + 1 := sum from i = 1; i <= n-1 = n(n+1)/2. The theoretical minimum number of trials of 3 connections to fully test all of these would then be n(n+1)/6. The theoretical minimum for six motes (16 connections) is 5, but I could only do it in 7 trials.