Wrist-Worn 3D Accelerometer Logger
Gathering data over a long time span with a high frequency
rate is a challenge, since it requires a device that (1) is
sufficiently small and light so that it can be worn comfortably,
(2) can store data internally over potentially longer periods for
later (offline) evaluation and (3) from which the raw sensor
data can be extracted, without preprocessing. It is hard to find
such devices that meet all requirements, especially since most
of these devices are closed-source and give only the possibility
of looking at preprocessed data via the company’s software or
by uploading data directly to their web server. Therefore, we
decided to rely on a custom-built prototype: the wrist-worn
data logger measures 3D acceleration samples with a default
sensitivity range of ±4g that are sampled at 100Hz, together
with the readings from an ambient light sensor (that could be
used later on in detection of dark environments). The on-board
accelerometer, the ADXL345 from Analog Devices, can be reconfigured from sensitivity ranges from ±2g up to ±16g and
supports sampling rates of several thousands of samples per
second. The whole unit fits in a plastic enclosure that protects
the module and is small enough to be worn comfortably on
the wrist and is attached with an elastic strap to it.
These sensor data are compressed and stored directly on the
embedded SD card for later retrieval via a USB port. As soon
as the logging device is thus attached to a host computer, its
data can be downloaded and the small battery can be recharged.
The 180mAh Li-Polymer battery lasts approximately two
weeks while continually logging at 100Hz, which can be
extended significantly by lowering the sampling rate (though
for this paper’s experiment purposes, we only needed to log
data for maximally two days). The designs, both hardware and
software, of this particular logger are open-source and available via http://www.ess.tu-darmstadt.de/hedgehog to support
reproduction of our experiments.
Study Participants
We gathered data from 42 sleeping lab patients aged
between 28 and 86 years, suffering from a variety of sleeping
disorders (though most were later diagnosed with primarily
Wrist-Worn 3D Accelerometer LoggerGathering data over a long time span with a high frequencyrate is a challenge, since it requires a device that (1) issufficiently small and light so that it can be worn comfortably,(2) can store data internally over potentially longer periods forlater (offline) evaluation and (3) from which the raw sensordata can be extracted, without preprocessing. It is hard to findsuch devices that meet all requirements, especially since mostof these devices are closed-source and give only the possibilityof looking at preprocessed data via the company’s software orby uploading data directly to their web server. Therefore, wedecided to rely on a custom-built prototype: the wrist-worndata logger measures 3D acceleration samples with a defaultsensitivity range of ±4g that are sampled at 100Hz, togetherwith the readings from an ambient light sensor (that could beused later on in detection of dark environments). The on-boardaccelerometer, the ADXL345 from Analog Devices, can be reconfigured from sensitivity ranges from ±2g up to ±16g andsupports sampling rates of several thousands of samples persecond. The whole unit fits in a plastic enclosure that protectsthe module and is small enough to be worn comfortably onthe wrist and is attached with an elastic strap to it.These sensor data are compressed and stored directly on theembedded SD card for later retrieval via a USB port. As soonas the logging device is thus attached to a host computer, itsdata can be downloaded and the small battery can be recharged.The 180mAh Li-Polymer battery lasts approximately twoweeks while continually logging at 100Hz, which can beextended significantly by lowering the sampling rate (thoughfor this paper’s experiment purposes, we only needed to logdata for maximally two days). The designs, both hardware andsoftware, of this particular logger are open-source and available via http://www.ess.tu-darmstadt.de/hedgehog to supportreproduction of our experiments.Study ParticipantsWe gathered data from 42 sleeping lab patients agedbetween 28 and 86 years, suffering from a variety of sleepingdisorders (though most were later diagnosed with primarily
การแปล กรุณารอสักครู่..
