Two tests of the platform were performed. The first occurred in November 2006, aboard a cargo ship travelling between Singapore and Kaohsiung, Taiwan. The results of this test were unsatisfactory, and although some sensors lasted for over a month in the field, many failed much sooner, due to excessive power consumption. Also, not all the novel sensors developed for the platform worked as expected.

After the tags returned from Asia, the bugs were fixed, and the tags sent out for another test in January 2007. This time, they were placed in a DHL express courier box, and sent by air between Boston and San Francisco. A current meter was connected to one of the tags to measure instantaneous current consumption, and the tag was able to maintain 23.7 uA current consumption while responding to external stimuli.

10-minute average power consumption of the tag during its journey between Boston and San Francisco. Periods of increased activity result in tag wakeups and therefore higher power consumption. The descriptions were taken from the courier service's package-tracking Web site.

The various sensors on the platform provide differing perspectives on the same phenomenon. From top to bottom: tilt switch, vibration dosimeter, and piezoelectric microphone.

In an effort to limit wakeups from aftershocks and increase battery life, the piezo microphone attached to the tag was subject to a dynamic threshold. As can be seen from the figure, the threshold increases with each wakeup, giving the tag respite from excessive stimulation.

The above two plots compare the stimuli recorded by the shock-sensor quasi-passive wakeup mechanism with dynamic thresholds of the CargoNet tag (top) to that of a tag sampling at 2 Hz (bottom).

As before, the top plot compares the events recorded by the quasi-wakeup system while the bottom shows the events sampled at 2 Hz. In this case, the sensor used is the piezoelectric microphone. The plots show good correspondence between the two tags, demonstrating that quasi-passive wakeup and dynamic thresholds are a valid substitute for periodic sampling.