Today Monica and Derick ran the first successful experiment (after two earlier attempts in yesterday and today) with 3 pucks in the MICU. The pucks broadcast a radio message each time one of the 3 selected manual soap dispensers was pressed or released. They also broadcast a heartbeat signal every 2 seconds. I parsed out the 88 press events and calculated the interarrival times, which are plotted for each dispenser in the graph below. The average inter-press time (lambda) was 400, 500 and 1167 seconds. The overlaid curves are exponential, which seems to be a good fit.
The data, when put all together, also has an exponential distribution with lambda = 597 seconds, as expected.
There was some interest in potential double-presses and what interval might be useful for the motes to separate double presses. Below is a sorted list of all the presses with inter-press times below 10 s. There are 10 less than 1 second and 3 between 1 and 5 seconds. There’s a gap between 3 and 5 seconds, which might be a convenient separation point for an empirical definition of the difference between a double press and two separate presses. However, if the data are truly exponential, there might not be a practical separation point, as the distribution itself would not be identifiable as bi-modal, as two exponential distributions can be combined into another exponential distribution.
0.268
0.310
0.321
0.322
0.325
0.350
0.379
0.399
0.477
0.625
1.525
2.432
3.171
5.119
5.373
6.634
8.914
9.919
