The shadow sensor: an electronic activity pattern sensor.

In their endeavor to measure time that individuals spend indoors, outdoors, and in transit, human activity pattern experts would be greatly assisted by a personal electronic sensor. This article reports on the design and pilot testing of an electronic device that measures activity patterns. The electronic sensor is small, unobstrusive, weighs about one pound, and records for 24 hr. The sensor identifies the microenvironment and the time spent in that microenvironment. Four types of experiments were performed during the pilot testing of the prototype: (1) quality control experiments; (2) other-directed experiments; (3) sensor vs. diary experiments; and (4) sensor vs. recall experiments. The prototype testing involved a total of 40 subjects. Quality control experiments were designed to test the accuracy of the sensor. Other-directed experiments were designed to test the veracity of test subjects. The subjects were told that the sensor was a pollutant-measuring device and were asked to record their activity patterns on a diary attached to the sensor. In the sensor vs. diary experiments the subjects were told the purpose of the sensor and were given a diary to record their activity patterns. In the sensor vs. recall experiments the subjects knew the objective of the electronic sensor, but they were not forewarned that they would be required to recall their activity patterns. The daily activity pattern difference across all microenvironments was the parameter used to quantify the discrepancy of information obtained from sensor and diary data. This activity pattern difference was more than three hours for the other-directed experiments, approximately two hr for both the sensor vs. diary and the sensor vs. recall experiments, and 18 min for the quality control experiments. The sensor provides an accurate description of activity patterns, and can be used to measure population activity patterns and/or evaluate the reliability of such information obtained by other methods.